Datasets:

salma-remyx

/

test_startup_advice_25k

like 0

pre Nuvo. Oh my lord pre Nuvo. I just went you know why I care about my health. This is my time to get as healthy as possible. And my bestie Chamath was talking about pre Nuvo spelt P R E N U V O. What is it? It's a full body MRI scan. And so many people have been telling me Oh my god, you know, I heard Chamath J. Cow's mentioning it and they said to basically save their life. That's what people have been telling us. So I had to try it for myself. I took out my credit card and I paid for it. And this is one of the most elegant experiences I've ever had. It's like going to an Amman hotel, like literally a six star hotel. You walk in, they greet you. You put on a nice little outfit. There's cookies, coffee. It's just, it's kind of like a spa, if I could say that. And then they screen you for over 500 conditions. cancers, aneurysms, it all takes less than 60 minutes. And it's no contrast and radiation free. This is proactive health care rather than reactive, then I get all the information I sat there with my wife, we went through it all. Listen, I'm in great shape, obviously, things are going fantastic. But there's a couple things going on my shoulder, my knee, and they said, Hey, this is something you should monitor. Renovo Thank you so much for making this easy to use service. They're in a ton of cities right now. They keep adding more cities. They love this week in startups. They love the all in pod. They said, Hey, we want to give a great offer to your listeners get $300 off at pre new vo.com slash twist pre and u vo.com slash twist and start taking care of yourself today. You know, it's really, you start thinking about the beauty of Silicon Valley. And this, the beauty of what we do, and I know it's criticized, and I know capitalism is imperfect. But there are a couple of things I've learned as I've gotten older, and we're old now. You're 54. I'm 52. I'm so young. I feel like I lost 30 pounds, you lost 50. Yeah. We lost 80 pounds. We're the best shape we've ever been. We're ready to work people. Don't put us out to pasture. We're talking about this earlier. I hope we keep that part of the show. But like Bill Gurley is like, and not a benchmark. And then, you know, Jeff Yang is still at red, red point, but maybe not as much. What's going on? These are people at the top of their game, they should be fully engaged. Stop with this early retirement for VCs, man. one of the beautiful things that happen is you start to become Obi Wan Jedi level Jedi, and you can start to see the patterns in the second decade, the third decade. And one of the beautiful patterns I've seen, and tell me if this tracks with you, is when things are aligned with the cap table properly. Everybody is incented on a milestone based funding environment that keeps employees, founders, and investors and hopefully the full course of investors align and rowing in the right direction. And when everybody's rowing in the right direction, and the founder says, Hey, I need help with something. And an angel investor and a series B and a series A and employees and past employees who've left but were given their stock options with a reasonable window and they were people were laid off, but they were still and a beautiful thing happens. Everybody's rooting for the company. And then in the situation you bring you bring forth. Now all of a sudden, somebody on the plane is like, this is not working for me. And they open the goddamn exit. You know, emergency exit and everyone the plane's like, we're going to Kauai, it's going to be great. We're going to Fiji. First, like no, I'm out. He starts fucking with the yoke, you know, no alignment. Maybe you could speak to the the beautiful alignment in Silicon Valley when it does work. Because what you've just described is when it doesn't work. Oh, I got the 30 million overhang. Sell this for 31 million. I don't give a crap what happens. I just need my 30 back. So I don't look stupid to my family office or partners. But talk about alignment and the beauty of this system that emerged over the last 50 years in Silicon Valley.

What's interesting when you talk about alignment because some of the advice that I give founders Jason and you'll know this through our own personal relationship and the times that we interact. I always say it's incredibly important as a CEO to help your board build personal relationships with each other. So I often recommend people do dinners. You do a board meeting, let's say from 1.30 to 4.30, have a little break, and then have a dinner from 6 to 9. So great. The reason is, I mean, obviously you can have conversations that came out of the board meeting. You can get to know the broader executive team. But, but having VC to VC relationships and or VC to independent director relationships, it's so critical because for the most part, like, you know, they say about flying, like they say, if you're a pilot, it's a really boring thing because you have hours and hours and hours of like boredom of nothing happening and short moments of complete panic. And that's like when you hit wind shears or bad turbulence or flying in the clouds, right? Yep. Engine goes out. Um, and that's what like boards are like, you know, we might spend four years being cheerleaders. And then all of a sudden it's like, holy, you know, we rip cord, like the doors flying open, whatever. And when people have personal relationships, they work through those difficulties together on the same team aligned. And when they don't, sometimes they work against each other.

Yeah, people, it's really interesting to me, what happens when and people have this thing, hey, the tide goes out, you see who's not wearing shorts, whatever. Okay, great. We know that one. The thing I find is when it's a storm, you know, like we're talking about and you're on the plane, you find out who can handle this kind of adversity, and who's not built for it. So everybody's high five and when it's up into the right, when all of a sudden you lose the top two lighthouse customers, you lost the CTO, you got six months of runway, you know, people start losing their minds, and they lose their composure. And then there's finger pointing. And then you and I are on boards, we've worked together, you know, countless companies. And it's great. Like, when people have been through it a number of times, I find now at this point, you tell me if this tracks with your experience, I have a certain sense of calm, when really intense things happen. It's almost like timestamps those for me, okay, the engines on fire, we're losing altitude, three people are screaming, okay, that's natural. Okay, what's the procedure here? Okay, we got to put the fire out, we got to get some altitude, and let's find a landing strip. And we all just start getting to work navigator finest landing strip. Okay, what's the procedure? And that is, I think, what time does for you, you know, when you've been through a couple of these, and you're like, you know what, and if this plane crashes, unlike an actual plane crash, we can shut it down in a classy way and start over. And what's your next best idea? Now, of course, the founder that's harder than for capital allocators who get many bets, but I find that over time, my blood pressure goes down. And in some ways, I almost look forward to the chaotic moments because it's an opportunity to be of true service, right?

Yeah i will say so as a starting point i think the best ceos and founders are calm in a crisis and they tend not to get to exuberant when things are going well and they tend not to get to panic when things are going and they always hit that point where they're going. I just did an interview jason with a journalist who wanted to ask me about ADHD you know i've talked publicly about having ADHD and he was asking me about crises and how i deal with crises and i said look. My brain is chaos you know. I'm my brain is always chaos. And I'm always trying to bring order to this chaos. And I think somehow with how my brain is wired, um, for whatever reason, in any sort of crisis, even a real world emergency crisis, I tend not to get too worked up. And I just go into problem solving mode, like what is the most important critical path issue we need to get done? Boom, boom, boom, boom. And so what I tell founders, Jason is, um, When you're deciding which VC to work with or which seed investors or angels to work with, reference your VCs for companies that didn't work, the ones that didn't work. Because if you ask anyone who invested in, I don't know, Stripe or Airbnb, where it was kind of up and to the right, everybody loves their investor when it's up and to the right. But in a crisis, how How did they turn up? I mean, you and I, like I can think of at least one board that we've been involved with before, where like early on in the company was like, oh, we're in a crisis. And I'm kind of stepping in saying, well, here's the steps and sequences to solve the crisis. We got through it. Then everything was up into the right again. Then we hit another bump in the road and what are we going to do about it? And being predictable in a crisis is a huge asset.

had a somebody I'm trying to remember who it was. They said the best VCs, the best board members act as shock absorbers. And the best founders act as shop shop shock absorbers for the company. Hey, you got hit. It's really hard. You got to take the punch. You got to take this, you know, bad beat, you got to get back up. And you got to set reality for the team. So let's talk about setting reality. last year or two, what has life been like setting reality and being a good board member being a good investor? We described, hey, in an up market, got to be thoughtful. Sometimes people are sometimes people aren't both sides of the table, the name of your amazing blog. Now let's talk about a down market. What have you been trying to do? What has life been like for Mark Zuster, upfront ventures, and the portfolio over the past 12 months?

So, thank you. Look, if you think about people like Sequoia writing RIP Goodtimes, if you think about David Sachs and the information that he put out when COVID first happened, like wartime CEO of Ben Horowitz, I think the service that VCs can provide to founders when the crisis period starts is this. Not just age and experience, but when you're dealing with 20, 50, 100 companies, you see the patterns before an individual entrepreneur. So an individual entrepreneur, let's say you raise money and you've got two and a half years capital, you may not quite realize just how permanent the capital markets have changed and how it's going to impact you in a valuation perspective. We realize it because we have three companies raising capital today. And so we're like, Hey, we see the trend. We know what life looks like for you in 15 months. And so we want to bring that forward. And so I'll just give you this Jason, which is if I take SAS companies, SAS companies were trading at in November of 2021. Okay. Not that long ago, 26 times next 12 month revenue. Twenty six times and again when you're a founder and let's say you know your first or second time founder you don't even know what that means like what do you mean twenty six times next twelve month revenue like why is that relevant to me i'm a startup. Can you say ok look. Eventually, you're going to either IPO or be purchased. And the person who's going to purchase you is going to have a public equity price, most likely, if not, you're going to IPO, and you're going to have to face public investors. And they're going to care about metrics. So one of the metrics they judge you by is either EBITDA, earnings before interest and taxation, and depreciation, or, or they're going to judge you by revenue. Okay. So let me give you some context. We hit 26, November, 2021. Today, it's at 6.2. That's how far public markets have come down. That's why late stage capital markets are going down so much. So you're like, holy 6.2. That's terrible. The market's surely going to double or triple, right? Cause we were at 26. Well, let me give you the data over the last 20 years. The average is 6.3 20 year average. The 10 year average was 17 and a half. So it had gotten out of whack. Um, and it's probably not going to stay at 6.2. So somewhere we think somewhere between six and nine times NTM is where it's going to settle out. And then late stage investors, if I think I'm going to exit at, let's say seven times NTM, I'm only going to pay 15 times NTM. If you're growing so astronomically quick that I pay 15 to get in, I sell at six and I still make money. Right. So David Sachs said publicly, I can't remember where I read it. He said, well, if your last round, you raised it a hundred times NTM next 12 month revenue, you've got to grow by seven X. You've got to be seven times bigger than you are today to raise it 15 times NTM. So what he's saying is cut your costs. lengthen your runway and make sure you have the time to grow to seven times bigger than you are today, or you're going to raise a down round. Okay. And I'm telling you 15 is still amazing relative to what markets are valued at, but, but that's also bigger than what the markets pay.

Yeah, it is. I guess there's this concept in behavioral psychology and cognitive, how we cognitively interpret the world of anchoring, what was something worth previously? Yeah. And maybe you could talk a little bit about how you've thought about thinking. Because when you get into this game, and you're making decisions, much like poker players, gamblers, or anybody taking big risks, unless the risk we take compared to the risk founders take is much different, but we have to take many risks over long periods of time on many different companies. So in fact, you really start to have to think and analyze your firm, and individuals, risk taking and how they make decisions. Maybe talk a little bit about how you think about and you've gotten better as a decision maker.

Um, so let me first talk about the conversation that I have with founders about alignment of interest between investor and founder. And then I can talk about my own decision framework. Um, so the conversation I normally have with people goes like this. You have nine months runway today as things stand, you're burning $1.2 million a month. I'm making all the data up. Um, you raised last round at make it up, uh, 120 posts. I can tell you that with market comps where they are today, your next round, if you were to raise today, would probably be at 30 to 40 post. Okay? So what's going to happen? Either you're going to run out of cash trying to raise and people are going to say, I don't really feel like doing a down round because I don't want to piss off all your investors. So I'll just wait for the next deal to come along. Um, or you're going to raise money and people are going to cram you down when they cram you down. They're going to look at me and say, Mark, are you writing a check? So I will write a check. So let's say that I own 20% of the company today. I can always write a check to fix my ownership, but if you're the founder, you can't write that check. So, well, you could theoretically, but most don't have the capital to do it. So I always say to people. I'm not telling you to cut your burn from 1.2 million a month to $250,000 a month to benefit me. No, it benefits you because if your runway can go from nine months to 23 months or 27 months, that gives you the time for two things, either to grow into your valuation or three things grow into your valuation. Number one, Or wait until there's a better capital market environment, number two. Or number three, when people go to fund you, when they make the investment decision, instead of burning 1.2 million, by then maybe you're burning 80,000 a month, right? Because you've kept your costs low, your revenue grows, your burn rate goes down. So instead of writing a $20 million check, they only have to write a $5 million check, and that's much easier to raise. So it's in your interest. Now you asked me about our decision framework. Let me tell you about our decision framework. Um, we've been around for 26 years. Okay. So I'm not the founder. Um, I've been running up front since 2011. It was founded my, by my dear friend and still colleague Eve sister on, um, in 1996. Um, and we did a rounds. You already said that our average check was probably four to $5 million in the past. Somewhere around 2015, 2016, when A and B rounds started getting so large, we had a choice to make. And the choice was, do we do one of three things, either write bigger checks, 10, 15, $20 million checks into A rounds. Do we B write the same size checks at the same stage, but own 10 or 11% instead of our target, which is 18 to 21%. Or do we still try to get 18 to 21%? Our strategy is defined by doing 40 deals where we own 18 to 21%. Our median ownership is 20% on our first check. And so how, how do you square the circle? What we decided was we would get religiously focused on just one investment thesis per partner. Okay. So we're eight people writing checks today. And each one has a swim lane. My swim lane for the last 11 years has been computer vision. That's why we work together on density. That's why we work together on fade. Like it's how do, how does the world get interpreted through cameras or lasers or sensor sensors or infrared? It's why I funded ring. It's why I funded Nanette, a baby camera. Um, but I have another partner and that partner is doing video game infrastructure. I have a partner doing healthcare. I have a partner doing FinTech. I have a partner who's doing cybersecurity. I have a partner who does AI and legit, sorry, not AI, robotics and logistics.

Which partner does robotics? I'm curious.

Um, so I actually have two partners who are doing it. My partner, Stuart Lander does it at growth and my partner, Kevin Zhang does it at seed and egg or seed mostly. Um, now by getting in our swim lane, What we started doing is saying, we still want to own what we want to own. We still want to write smaller checks, three and a half million dollar checks, maybe four, maybe five, but on average, three and a half. It meant that we had to move earlier in the cycle. Instead of waiting until you have product market fit or, you know, customer references or, you know, whatever, I have to back you. Usually when you finish the product, But haven't yet hit revenue, because if I wait until you have some element of product market fit, someone else was lined up with a $20 million check and I wouldn't win that deal. So we moved to fund companies earlier and faster to maintain our ownership and to get in to the deals that we wanted to get in.

Running a startup is like being a small market team and you're trying to compete against somebody with unlimited resources like the Yankees and the Dodgers. Well, if you've seen Moneyball, one of my favorite films, you know that using data correctly can help you compete against those big incumbents who have seemingly unlimited resources. One thing that startups haven't really had access to until now is detailed scenario planning. This is stuff that like big companies get to do. There really haven't been tools that are affordable or elegant enough for us in the startup crowd, where you can do this easily with org space, O-R-G-S-P-A-C-E. It's people, software, for software, people. Basically, it lets you create plans for deploying the capital that you just raised and money you're making, and then adjusting headcount based on different future scenarios. For example, what if you raise your Series A? You get 10 million in the coffers. right? And you got to deploy that. What if you can't raise right now and you got to make it work with your $3 million seed round? What if your revenue goes up 20-30% next month with org space, you're going to be able to plan for hyper growth, plan for rifts and everything in between. Things like cost, skills, DEI, all that is in context. So you understand the impact of your decisions. This is the thoughtful way to do it, folks, twist listeners get $2,000 in credits on org spaces, pro plans with a 30 day free trial at org space.io slash twist, org or g space as p a c.io slash twist get those $2,000 in credits. It's really interesting with this cognitive biases we start to have in the signaling. What signaling do you have right now around AI, generative AI? the meme of the moment, the focus of the moment. And what signaling did you have around web 3.0 and crypto? Did you get it right? How do you look at it going backwards? Do you did you do you regret things? Or do you felt like you had a good decision making process?

And then here's where age Yeah, of course, of course, sorry to jump in. This is where age helps. Okay. So the advantage of youth is their peer group are the people creating companies for the most part, right? Like some people create companies at 48 or 52, but for the most part, it's younger people. And when you went to Stanford or Yale or Princeton together or Michigan State or Washington University, wherever you went, but When you went together, and that was your peer group, and then you all got jobs together at Google or Facebook or Stripe or Dropbox, right? You have meaningful relationships, and they're more likely to want to raise money from you. So we have gone out and hired younger partners. I have Aditi Maliwal. She's much younger than I am. Kevin Zhang is much younger than I am. Kobe Fuller is much younger than I am. So they run in different circles and have different networks and crowds. And that's a huge benefit to me. But I'll tell you what happened in the crypto craze. I said to people, that's not our swim lane. If anybody wants to drop what you are working on and go super deep in crypto, Then we can have a discussion, but until such time, we're, we're just going to skip that. It's okay. If Chris Dixon and Fred Wilson make a ton of money on that because they were early. So to be right about venture, to be good at venture, I think you need to have three things, right? You have to believe in a trend that's going to happen in three to five years that most people don't see yet. Okay. You need to be correct about the timing of that. If it's eight to 10 years, as you know, Jason being too early is the same as being wrong. Yep, if you're reading about it today, you fucking missed it. Yep It's okay The next trend is coming but if you if you pushed your whole pile in if I take your poker analogy if I push my whole pile in on crypto in 2020 and I wasn't doing it in 2013 14 and 15 like chances are you know, like it was, you know, your old pal tony shea who said like um that his thesis in Las Vegas was not to be the best poker player at the table, but to sit at tables with the worst poker player. Yeah, for sure. Or if you don't know who the sucker at the table is, it's you. It's you, for sure. So if you push your pile in in 2020, you're probably hurting right now. So in 2020, 19, 20, 21, I was under a lot of pressure from some of my colleagues saying, we need to pay more. It needs to be 40 pre, we need to do NFTs. We need to do crypto. And I said, well, I think you missed that trend. And there's some great people who caught it and understand it better than you do. If you want to drop everything and go deep, I'm here. Nobody wanted to do that. Right. So we didn't do crypto. We skipped that.

Why be tourists? I mean, why come in last? You're the sucker at the game. You know, I looked at it and I made the assessment with crypto. that I like to talk to customers and or talk about the product and how it was constructed. And I consistently got told have fun staying poor. You don't get it. Okay, boomer. And I said, Well, this white paper has spelling errors in it. Anybody could have written it. I'm sorry, I'm not the right investor for you. Because I like to look at the product and talk to you about why you put the buttons in certain places, the workflow and, and how it's gonna sort of hit customers. And you can't explain that.

So find another investor are some really smart young investors who just were passionate about it. mags being one Gabby goldberg being another there's money to be made in web three we know that this distributed infrastructure will produce some interesting things. But i'm not the expert in it and we chose not to go along on the other hand we get really big. On computational biology and on healthcare and we started doing that like eight nine years ago. And we have some really interesting companies in the category now and everyone's like what are you doing so but let me say that so first of all you need to be right about the trend second you need to be right about the timing and third you need to bring back the right team so i found Uh, this great trend. I had just moved from Europe and Japan back to the United States. I lived in, um, Europe and Japan, as you know, came back to the U S and I said, I don't understand why there aren't text messaging companies. There's going to be something big in test messaging, text messaging. So I went out and I met a bunch of teams and I met a fantastic team based in LA. They had built and sold their first company for $580 million. And in mobile game 1.0, they were the winner in the category. They built a company called jam that. And Mitch Lasky, who went on to benchmark was the CEO of that company. He didn't found it, but he was the CEO and he's incredible. And actually the people who built it were incredible. And they built a company called text plus, and it was up into the right and there were four or five up into the right. But the problem is there was one winner. And that winner is called what's up? Yeah. And I could go back and I could have what they took. I don't know 99.9% of the value. And I could go back and say, why didn't we win? And we could second guess and all the things. It was a great team that we backed. We just didn't end up backing the winner. And that's the hard thing about being a venture. You have to be right about the trend. You have to be right about the timing.

And you have to back the right I mean, look at Uber, right? There was there was lift inside car. And I met both those companies before Uber. I knew the founder of Uber longer, but I took a sidecar after I'd invested in Uber and I was in a sheer panic. I mean, I called Travis and I was like, I just got an in sidecar and let me pick the price I want to go Palo Alto. And he's like, well, take a couple rides. Let me know what you think. And then I took the Lyft ride sharing and I was in a total panic because we were only doing Lincoln town cars. And I said, Well, with sidecar, I offered somebody $25 to take me from San Francisco to Santo Road. And Uber would have been 75 lift was like 60 or 50. And this guy did it for 25. But I felt like I was gonna die because he was leaning back so far in the chair that I was having like a face to face conversation with them. car smelled like weed. It was like scary. I was like, so that one's definitely not gonna win. People should not be auctioning off the lowest price for a car doesn't feel safe. But this lift thing, with the Prius, I kind of like the Prius better than being in the SUV, because I don't feel as douchey. And he's like, we've already got it built. I said, I mean, I'm telling anybody it's built, we're just gonna press the button at some point, we're just going through some, you know, scenarios or whatever. Okay, great, back the right guy. You don't want to go to war with that guy. And then I subsequently met the door dash team. And they said, there, they would wake up in cold sweats. Travis, Travis, you know, and it just is the nature of it. You're right. Like if you that the some founders are so transcendent, they're just so good at winning. that yeah, you can get everything else right. The great thing about being a capital allocator, you need only hit one per fund. Is that true with the power law and explain the power law to people who maybe are new to venture capital, and or new to startups, because that does affect and you've written about this a lot of times again, the blog is both sides of the table. So you can understand both sides table. What do founders need to understand about the power law and how that impacts behavior of venture capitalists in terms of you know, outcomes?

So let's say you raise a $300 million fund and let's say that you're investing in 40 companies. And let's say that you own 20% of a company, you gave them $3 million and then they turn around and get offered to be acquired like quickly. And you're like, dude, I just made you three times your money in three months. Okay. that $3 million, three times my money is $9 million. So I make a gain of $6 million in even in three months is a total loss for me. Because 9 million, I got a return at least 900 million. So you've returned 1% of the minimum expectations that my investors have. So if you think of 40 shots on goal, I just wasted one of them. You're thinking I got a huge victory for you. I got three X and I'm like, that's a waste. No, that doesn't mean that I'm not happy for you or you know, whatever. I don't move on in life. Like we're not dicks. Like if that's the thing you want to do, but I'm trying to back people. who aren't looking for a quick exit, but really are driven by some bigger mission. Like I don't know what drives Elon. You obviously know better than I do, but something more than money drives that guy. And you want to find somebody who they're driven by some other thing. That's why I always talk about wanting to back passionate entrepreneurs that are like really driven by a mission. Like we talk about missionaries versus mercenaries and the missionaries. are the ones who can get through crises. Yeah. And so like the math just doesn't work. So when you think about a $300 million fund, and I write 40 checks, what we know across more than 11 funds, is that, and you know, it's the power law. 20% of our investments, usually it's less than 20% return 80% of our total returns. So in a four $300 million fund, it's usually five or six deals that return 80%. And it's usually one or two that disproportionately return the fund.

Yeah, this is very important, I think, for founders to understand. And it's, it's great that the the missionary approach as opposed to the mercenary approach. What's really interesting about that is we're trying to get an outlier, we need you to swing for the fences. mercenaries, they don't swing for the fences. If they get that 40 million 80 million eggs that they own 60% of the company, hey, it's life changing money. They're gonna pull the ripcord because they can imagine themselves not working on that mission. because they did it for the money. And the challenge in the alignment, and this is one of the few times where there's like a misalignment in venture is, yeah, sometimes that early sale comes, and I had it happen two or three times, where I begged the founder, keep going. And the problem was, in the early part of our careers, the concept of secondary shares did not exist. So the founder was going to get a zero or whatever early exit they could get. And when the early exit comes, well, you know what, getting $25 million life changing money, you say yourself like, Hey, I'm happy for you. But it's you're sad for you. The mercenary versus missionary is a great way to look at it. But maybe you could unpack why secondary kind of corrected this one of the few misalignments in venture, which is hey, the quick exit is really good for the founder. And it's a disaster for the VCs.

Well, I don't want to pat myself too much on the back, Jason, but I think I was the first VC to publicly speak out in advocacy for founders on this. Um, and this is already, I think I wrote my blog on this in like 2008. Um, it used to be the mantra of our industry that founders can't sell secondary until the whole company is sold because every dollar, all the capital should go towards the success of that company. And the reason I spoke out was I had, like you, you know, in your founder days, I had been that guy who had a startup and my VCs all lived in Atherton or the equivalent in London, Belgravia. And, you know, I had my wife saying, why are you working for this small amount of money? I was earning a lot of money before I did a startup. And, you know, I had a little kid, I had a second on the way. And, you know, we're living in a shitty house and it's a struggle. And I'm like, you know, if I could just make a little bit of money, that would change how my wife felt about her own sense of security and why I don't come home on weekends or, you know, why I'm flying on weekends to meetings and staying late every night. Yeah. So I coined this term, feed the family money. And I started talking about feed the family money. And if I could just get founders enough where their family Metaphorically right lecture you take that pressure off you and the hard thing is when is it okay it's obviously not okay before your company is a success so when you have your company has some level of success that you can feed the family. Or put aside money for your retirement or for purchasing your first ever house or whatever it is. Once you've hit that little bit of success, what it means is we have aligned interest now because you're not playing small ball. And now you're saying, I really want to do something extraordinary. And I still believe that today, but I will tell you, secondaries are really hard to come by these days.

Yeah, secondaries. I haven't seen many of those recently. Because people are trying to, I guess, fix their portfolio construction. But I you know, I think when people figure out who the winners are, there'll be a, hey, this company has a good cap table companies growing short would love to add to my position in it. Because I got to put the money somewhere. And this feels like a safe haven for the money. And I've already vetted the company.

The hardest thing that people have with secondaries today as we sit in 2023, early 23, is how to value them. Because if the public markets are paying 6x forward for a SaaS company, if your last round was at 18x or 24x, and you don't want to sell at a huge discount to that, You know why why would an investor go in and pay 12 times or 13 times for your supposed discount. in a stock that is not top of the preference stack, and doesn't have downside protection. I'm not arguing against founders. I'm just telling you how investors think these days and why it's harder to come by.

Yeah, it is one of these systems where it did get abused. I remember seeing some deals where the VCs were being selected by the founders. by which vc was offering the most secondary this felt like a huge conflict of interest i said hey maybe we should separate these two things just for hygiene raise around that we do the secondary are we picking the partner for this next round.

Based upon how i was putting the sweetener it happens in every cycle and every boom market. Again i don't really blame market participants like the money in the temptations were too big but it really drove some absurd behaviors and what people often don't understand jason is that. Sometimes it's a founder CEO at odds with their company and the company doesn't even realize it. Right. Like, so if you take $15 million off the table in a company, that's not profitable, not really guaranteed success. And the CEO takes 15 off rank and file do not. And you ended up picking the wrong investor just so you could pocket your 15 rank and file. Don't even understand what happened to them.

They're not in the loop. Should there be more transparency for the rank and file? Should any of the secondaries be pari parsu? If the founders participate, everybody else should participate. What's your position on it?

Everybody looks at it differently, but really, I think the easiest way to make it work is if a CEO wants to sell, they probably should offer it to rank and file. Now you could, you can slice and dice your data differently, Jason. You could say, for example, anyone who's been at the company for four years or more can sell. You can say, I'm making up the date. That's like an individual decision. It could be. Um, you can sell up to 15% of your position, but I, the CEO, I'm not going to sell more than 15% of my position either. So it could be relative. Um, like it's okay to have some slices and dices because like, you don't necessarily have someone who's joined three months ago cashing out. Right. They haven't really contributed to the success. That makes a lot of sense to me, but there ought to be a structure.

It should be an orderly process. I think this double dealing ice, we, we know this famous instances of this, but you know, like some founder. Yeah, they know somebody who's a, you know, some crazy billionaire, they sell, you know, a third of their shares before the IPO. Everybody's locked up. IPO price was a little too high. The price before the IPO was a little too high. And they get to clear a third of their position. And then everybody hates them. So you got to really be thoughtful about this. Because again, alignment does matter. You need to everybody aligned. And if everybody's not, and we have started having these bad feelings, it gets toxic. Maybe you could give us an example. I know we're getting close to the end here. some examples of things that were company killers, and some things that were, hey, really profiles and courage, things that really got the company aligned, and you can composite it, obviously, unless makes everybody look great. But some thoughtfulness around things you've seen that have that have crashed the plane that didn't need to, and what you took from it.

Well mostly what i should tell you jason is my philosophy and you know going back to psychology where you were like earlier in the episode which is. There's no such thing as a good or bad bc the pool of vcs and their behavior. in is just maps what there is in the general population, there's amazing VCs, who are really hardworking, thoughtful, earnest people not in it for the cash, that are show up and be dependable. And there's apples. And the same is true of founders. Like there's not a larger proportion of altruistic founders, then there are VCs, then there are investment bankers, then there are lawyers, like we're all just human. And so I will tell you that There have been some bad behaviors for CEOs and these things never get public. Um, and in a bull market, what we saw a lot of CEOs doing is pushing really hard for personal top up in their shares. So let's say you did around and they come back to the board and they say, I want to own 6% more of the company. And the board might say. But you already own 18% and the rank and file each own less than a half a percent. So we'll take dilution, but it really should go to rank and file. And by the way, you already sold $10 million a secondary in the last round, and that's why you own 18 and not 24% or whatever. But in a booming market where they had a million options, what happened was founders who found someone who was willing to let them top up their personal shares. And they didn't always look after their staff. Now, again, I don't want to suggest that there's not VCs who do self-dealing. Uh, there, there are. There's not good and bad.

How do you start to see the predatory terms, the three Xs, the two Xs, the cram downs?

Is that happening now? Oh yeah. I just read a report. Oh yeah. I just read last night, um, a data set that was put out that, um, maps across a whole bunch of terms, whether they're founder friendly or investor friendly and the nadir, the low point of investor friendly or the peak of founder friendly. No surprise was 2019 to 2021 for the last 20 years. And it's a sharp trend up towards investor friendly right now. So I wouldn't say three X participating preferred, but you are seeing participating preferred coming back.

Okay. I get my money and then I get my percent ownership. You get a double dip.

You're seeing dividends going up. You're seeing, um, it, you might see a 1.4x liquidation preference. How does the dividend work?

You're not actually getting the cash back, you're getting what?

How does the dividend work? Well, you can have things like a PIT, you can have things like a PIC, which is a payment in kind. So it basically says you accumulate a dividend, and then you get it in equity later. So it's just a way of getting more ownership. So you feel like you sold 18% of your company, but one day when you sell it, you really sold 23% of your company because a dividend builds up that gets paid in kind.

like the investor put in 10 million, they get 6% dividend. Every year they get another $600,000 on that 10 million.

Yeah, it gets put on top and they and they can put it they can either have it stack up onto their liquidation preference. So that's downside protection. So if my liquidation preference builds, which they typically do, they typically increase over time. Um, or I can actually have a payment in kind, which means that I get equity cumulatively and I get it paid later at that, um, whenever the evaluation was at that time. You also see things like full ratchets that come to play, um, in a market like this. Um, and honestly, the, the pendulum hasn't swung fully Jason, right? Like there's still an oversupply of capital. So the terms are relatively favorable to founders, but it is changing quickly.

I think Yeah, if you're a founder, the best advice is to build as much runway. So you have as many options as possible. A founder with 24 months of runway right now can turn down deals a founder with five months of runway, six months of runway, they reasonably can't turn down a deal, which means we know how that dynamics going to go. going to go down to the wire, and you're going to wind up taking a really bad deal. So be thoughtful. Mark, as always, our conversations. So candid and so insightful. It's great to work with you. If you're a founder, read Mark's blog, both sides of the table. And if you're a seed founder, you're not going to do much better than Mark in terms of somebody who's going to work hard for you. So go ahead and pitch him your company. The best way to do that is

Well, uh, I, I still believe the best way to get ahold of us is to get introduced by somebody we know, which is usually the founders that we've backed.

Oh, that's the best way for sure.

I do. I do read email, but like, I get so many, like you, so many inbound that how do you know which one of those things to really focus your time and energy? Um, so, you know, getting a friendly intro and, and like, I know people get outraged by this. Like, why should I have to get an intro? But actually the skill set that it takes to get an intro to a VC, it's pretty easy to get introduced to a VC. Like we are predisposed to want to meet people, but that's the same skill set you're going to need to sell your product to enterprise clients, to get journalists to write about you, to do business development deals, to persuade people to join your company. So it is a bit of a test. That's the first test is how do you get access and what do you do with it? Yeah.

That is one of the, it's one of the first tests. I have one test right before it that I always tell people is the first test. My first test is, can you get a co-founder? I get a lot of emails because I tend to invest a little bit before you. Um, we overlap, obviously, but they, you know, I can't find a co founder, I might fail the first test, can't get an intro, failed the second test.

The first two things I look for in any company. is cadence of recruiting. Obviously, quality matters, but cadence of recruiting and cadence of shipping product, because people who can't hire and can't either because you don't have access or you don't dedicate time, or you're just slow and decision making, they're going to be like that for the next 10 years. And people who can't ship product regularly, it's a bad pattern.

Yeah. I agree. The second one is my first. I look at that product velocity. I just love product velocity. And you describe this in probably your most famous blog post, invest in lines, not dots. Explain to people as we end here. This philosophy came up with and do you still believe in it?

Or have you edited in any way, completely believe in it. And I think it's a two way process. Okay. So the idea is think x axis, y axis, where x axis is time, and y axis is performance. however you want to measure performance, okay? When I meet you, you're a dot. We either had a great meeting or a bad meeting. You either were on a high because you just got a bunch of wins or you're on a low because you had a bunch of losses or whatever, but you're a dot. And the next time I meet you, that might be up to the right or it might be down a little bit. But over time that I connect those dots and I see a pattern like you got kicked in the nuts or, you know, kicked in the shins or whatever metaphor we want to use, I guess, gender neutral, you got kicked in the shins. You got knocked on your ass. You got knocked on your ass and how did you get back up? And did you dust yourself off? Were you resilient? Did you have good followup? Are you good at recruiting? Can I see what's changed in your product? Can I see what's changed in your forecast? Did you get pressed? Did you, whatever. And over time, if I've met you four or five times, I start to detect a pattern of what it's like to work with you. And the same is true of a VC. They might show up their first day and be super charming. But then they never really follow up on things. And, you know, if they don't follow up when they are supposed to, uh, be in the courting phase of you, well, that might mean they're not interested. That's true. Um, but if they say they're going to do it and they don't do it, imagine what it's going to be like when they're on your board. So like it's a two way street and like someone who is thoughtful in the first meeting with you and has all sorts of ideas, but the second meeting, they don't remember even really what they talked about the first time. Yeah, like that's probably what it's gonna be like when they're on your board.

So please tell me what you're working on with these brain implants.

Thanks, Molly. Thanks for having me. It's a pleasure to be here.

Right. How common are these disorders?

Well, we're working on brain computer interfaces is the general term for the technology that we're building. And yes, they are a form of brain implant. And they're designed to connect the brain directly to computer systems as ways of helping to treat some forms of neurologic disorder that are currently basically untreatable. And those include things like certain forms of paralysis, stroke, traumatic brain injury. forms of disorder in which the brain can think but the body can't act. And, and ring computer interfaces are designed to enable a direct communication between the brain and a computer bypassing the part of the body that isn't able to act in order to reconnect the brain to the digital world.

Right. I do not mean this in any way to sound insensitive, but I just went through a version of this with my dog where he was losing the use of his back legs and it was simply, you know, the vet was like, his brain is not telling his legs to work and it was super terrible. I want to kind of go through the history here because you're at Precision now, you were at Neuralink before that, but I want to go all the way back to your background and what, like, what is the origin story here? What got you interested in this? And more importantly, what was the moment when you realized that this could be possible?

Well, there are definitely millions of patients, millions of people in the United States alone living with some form of paralysis from spinal cord injury or other disorders. Almost everybody knows somebody.

That's amazing. So you were in the science fiction part of it, where you imagined a future that could be possible. And then that has become true within your lifetime, which is amazing.

Well, well, that's it. That's a great way to start. And, you know, no one ever really begins something completely de novo, right? And I come from a family of doctors and engineers. And in a way, I guess I've been working on on this my entire life. My dad is a neurologist who specializes in electrophysiology, which is the electrical aspects of the way the brain and nervous system work. And my grandfather was an electrical engineer, a radio operator in the Second World War. Actually, my father trained to be an electrical engineer and was exposed to the very earliest forms of artificial intelligence. And in a way, that was how he made the transition to becoming a doctor. So, I grew up with electrophysiology and clinical neuroscience as part of the everyday. And by the time I was about 20, finishing college, the most interesting thing in the world to me was what was just becoming possible or just seeming to become possible at that time in the late 90s, early 2000s, which was the notion that even though for a long time it was possible to interface electrically from a scientific and clinical perspective, with the nervous system. And in fact, all through the 20th century, research neuroscientists and doctors had been using the electrical properties of the brain and nerves to diagnose and treat disease and to study the nervous system. the electrical nature of the brain and nervous system is kind of what makes it special in the human body. But it was not possible to do that in a kind of high bandwidth way until the very end of the 20th century. And what I mean by that is that you could maybe record from a small number of nerves or a small number of nerve cells at a time using specialty hardware. until the very end of the 20th century, it became possible all of a sudden, through some breakthroughs that maybe we'll talk about later, to record from many, many neurons at a time. And that change in the bandwidth of our ability to interface with the brain and nervous system made the current generation of brain-computer interfaces possible. And I saw that happening and to me it seemed incredible and I basically spent the rest of my life, have spent the rest of my life working in that space. And early on, very early on, scientists, neuroscientists, understood that it might be possible to restore function to paralyzed patients, amputees, spinal cord injury patients, even in some cases blind patients. And so, the promise of the technology has been around for quite some time, maybe 20 years now. But it wasn't until about the late 20-teens that there was kind of a general consensus that it was ready to emerge from academia into the real tech world to really translate what had been proven in academic settings into clinical reality. But that was something I had wanted to do for basically a long, long time. I and many others and that's what we're set to do at Precision.

Start with, if you wouldn't mind actually, give us a primer on the electrical nature of the brain. For people who may not be familiar, you know, you mentioned this is the thing that makes the brain and the nervous system special. If you wouldn't mind, just give us that kind of like 100 level, why that is the case and why there became this idea that you could potentially tap into that in some way.

I guess I guess you could say that. Yeah. I mean, in a way, that's how a lot of good science gets done, right? The science, the science fiction of the prior generation inspires the next generation to try to turn science fiction into into fact.

That's fascinating. So you needed this combination of the ability to manufacture an implant that is smaller than a human hair, as I understand it, combined with the ability to have adaptive compute, basically, on the other end to say, okay, I'm taking these signals, I can decode them and I can put it's not a brute force problem, you can't apply the same you know, zap to everything.

Sure. Well, neurons, which are the cell in the brain that are responsible for conscious thought and communication, and many of the functions of the brain and body that we think of as making us human neurons, communicate with one another using electrical impulses. And those electrical neurons are tiny. They're about if you put them side to side, the body of a neuron, you might be able to put 20 in the space of a millimeter. So, they're very small and the electrical signals that they produce are also tiny. But if you think about it, if you make an analogy to kind of sound, I think sometimes that's an easy way to think about it. Instead of thinking about electricity, think about sound and think about the ways of interfacing with neurons and the brain kind of like listening to the brain instead of electrical, just because I think that's an easier way to associate with what's going on. So, if we talk about in terms of listening, electrically interfacing with the brain is kind of like in some ways listening to or speaking to the brain. And interfacing with a neuron or with groups of neurons is kind of like building a tiny little microphone that we bring just up just up close to these tiny little neurons, or that we place within a group of neurons. And we try to listen to their chatter. We listen to the way they speak to one another. And, and there are distinct patterns of electrical activity. that we can hear and make sense of and we call that decoding. So, the electrical chatter of single neurons and groups of neurons is a language. Every brain speaks a little bit differently. The problem of learning how to interpret what the electrical signals are in the brain, what they mean is a little bit different from individual to individual. And so, that's kind of where the artificial intelligence angle comes from. And that was part of the technological change that occurred in the field of brain-computer interfaces as we moved from the 90s into the 2000s and the 2018s. the ability to the material science that allowed us to record from many neurons, or large groups of neurons at once coincided with increases in computational power and sophistication that allowed us to decode what those conversations among neurons meant. And so that in a way that that confluence of technological paradigm shifts, as what has given is what has given rise to a very powerful new technology, which is the brain computer interface.

So now let's talk about thank you for that. That was an incredible primer. Also, you're a total poet. The analogies are perfect. Um, So then talk to me about the process of forming companies around this. So you started by founding or co-founding or partly founding Neuralink around this technology. Tell me about sort of the history there, how you became involved with Neuralink.

I was one of the one of the eight co-founding team members at Neuralink back in 2016, 2017.

Okay. And then explain the talk to me about the innovations that were being worked on there.

Yeah. I mean, what I'll say is that I kind of mentioned historically the way the field has developed and I can maybe take things back in history a little bit further just to give some context. So, you know, the origin in a way which just to give some deep perspective is, you know, the 20th century was when biology really discovered the electrical nature of communication in the nervous system. And that's what we were talking about earlier. And the way that scientists and doctors interrogated the brain and nervous system was with electrodes. And an electrode is just a device sometimes it is a wire, sometimes it is another kind of conductive material that allows you to either touch or come in close proximity with the part of the nervous system, could be the brain, could be a peripheral nerve, could be a spinal cord that's generating electrical signals. And up until the late 20th century, there was no real standardized manufacturing process for manufacturing those electrodes. And neither was there a really standardized way of processing the signals. And so it's not like in audio engineering, you know, there was a, there was a whole industry that standardizes the manufacture of microphones and, you know, equipment for amplifying and processing sound and filtering it and recording it, you know, there are standards and, and known equipment that you can buy that didn't really exist. very much until the late 20th century. And what did exist was relatively large scale devices. And let's say on the scale of fractions of a millimeter, okay? So, that's large in neuroscience. Lots of human hairs, yeah. A ponytail. So, when I say large, it's still relative. And then in the late 80s, early 90s, what I think was really a turning point for neuroscience and for what became the field of brain-computer interfaces was the development of a device called the, what's now called the Utah electrode array. And that was a micro array of tiny little electrodes spaced at a fraction of a millimeter from a part, approximately 100 of them. And they were made using the same manufacturing process, same microfabrication process that is used to manufacture microchips. Okay, so you could make this electrode array of 96 or 100 electrodes, you could make many of them, they would all be exactly the same. And you could, you could give them to researchers to use, and everyone would be recording using a standardized microfabricated device. And looking back on it, I think that was the moment when Moore's law arrived in neuroscience. Wow. And remind us, I'm sorry, what year this was? The late 80s, early 90s was when Richard Norman developed the Utah Electrode Array. Got it. So then you had the chip. Well, that was the beginning. It doesn't happen all of a sudden like that. But you went from artisanal manufacturer of electrodes to a standardized microfabricated manufacturing technique, and one that allowed high performance microelectronics to interface with the electrodes themselves. And, you know, we are all familiar with the, you know, the kind of concept of Moore's law, the notion that some scaling property could be applied to the technology. And the scalability of microelectronics is what is one aspect of what has powered the the revolution in computing that began in the last century and continues today. That scaling paradigm is essential, and the ability to connect the electronics to the end effector was essential. So that only arrived in neuroscience in basically, let's say 1990. And that paradigm was pushed, you know, into the early 2000s. Remember that, you know, high-performance computing, as we think about it today, the kind of things that enable modern artificial intelligence applications that didn't exist until the mid-2010s. So things that we kind of take for granted, high-performance computing applications that we take for granted, even in things like image processing, you know, did not exist at that time.

Ironically, it also required sort of a new chip architecture, right? The shift from CPUs to GPUs.

That's true. Correct. So GPUs did not exist at that time. So the early interfacing of software with the new generation of microelectrodes was all CPU-based computing, which worked just fine for tens of electrodes. And it was kind of strained at the hundreds of electrodes level. But now, you know, in the 20-teens, computing was able to catch up. And concurrent with all of that, you know, there was a generation, I include myself in this, of, you know, engineers in training, masters and doctoral level engineers who kind of cut our teeth on circuit architecture, algorithm design for how to interface with the signals coming off of these sorts of electrodes. And by the early 20 teens, I think there was a general consensus that most of the major science problems or many of the major science problems, even some of the engineering problems had basically been solved. In other words, how to how to build back-end electronics and how to encode software that would make sense of many simultaneous electrical signals coming out of the brain in a way that we could understand what the brain was trying to tell the arm or the leg or the mouth to speak. And at that point, there was a sense that to take the next step, to really transition academic science into clinical reality that would benefit people. it was time to move into a commercial and industrial setting out of the lab. And I want to give you one other piece of context, which is that, you know, think about the historical backdrop of all this work as it was happening in the United States. you know, from the early 2000s, 20 teens, we were seeing a lot of, you know, wounded young people coming back from Iraq and Afghanistan and the National Science Foundation and DARPA and other funding agencies. had a strong mandate to try to do whatever was technologically possible to take care of these motivated young people who had been serving our country. And so, that led to a tremendous amount of attention being paid to advanced prosthetics development. Not all of that was neuroprosthetics. the work took many forms, but certainly that motivation catalyzed a lot of tremendously important work. And so the government funding agencies invested heavily in the development of this technology. But at some point, it became clear that that that form of investment and the timescales that were required to secure the government grants and do the work in an academic setting, the funding wasn't enough and the timescales were too long. And in order to really translate the advances that had been developed in academia into clinical reality, it was time to move into a commercial setting. And so, in 2016, a couple of major efforts shifted attention from the academic to the commercial setting. Neuralink was one of them. Facebook had an initiative around brain-computer interfaces. A company called Kernel also was started around that time. Each of these entities put a huge, what at the time seemed like a huge amount of capital behind moving people and resources from academia into a more commercial research and development setting to build the brain computer interfaces that would actually go into the clinic. And that has given rise to a small ecosystem of advanced startup companies and a tremendous amount of talent being brought to bear in the field. And I would say that that's the best thing that happened out of Neuralink, is that engineering talent has been really focused on what I think is a tremendously important problem for our generation. And the more the more talent we bring into the field, the better.

So it was sort of it was sort of a beacon, and it attracted all of this best and brightest, and you could sort of go off and found related companies.

hiring freelancers and doing that on a project based work is a brilliant way for you to grow your startup sustainably right you can't just hire everybody in every little vertical and listen there is a ton of top talent right now out there looking for work through all the layoffs in tech you know that So you need to check out Contra, C-O-N-T-R-A. Contra is a commission-free marketplace for freelance and independent creators. So all that money that's going back and forth between you and your freelancers, it's not getting taken by some marketplace, no. There's no percentage-based upcharge when you do hire somebody. And they do all the vetting, they find the best people. On the other side of the marketplace, hey, if you're one of these laid-off tech workers and you've got tons of skill, sign up for Contra. It's an amazing platform for you. And remember, like I said above, creators on Contra keep 100% of what they make. There's no fees. They specialize in design, engineering, social media, video, writing, and of course AI. This is a really easy way for you to get great talent and to do it quickly if you need project-based work. You need to check out Contra. It's that easy. And you know what? The best thing about freelancers is you only spend what you need to spend. You might have a really important social media project, but it's only for six months of the year. Or you need some videos, but you only need 10 of them, not 100 of them. They're going to do it fast. They're going to do it right. So here's your call to action. I can't believe it. $500 off your first hire at Contra.com slash twist. That's right. Five crisp hundies waiting for you at Contra, C-O-N-T-R-A.com slash twist.

I wonder, talk to me a little bit more about that. I think we're all familiar with the kind of academic valley of death that can occur with R&D where it doesn't become commercialized, or there is just not quite enough investment, right? It's like you can get here, but you can never make the final leap. And yet, there is also that question about what happens if you attract all of the best and brightest researchers and scientists to private industry. that if that doesn't work, then you've left universities unable to continue this research to like, how do you tackle that tension behind doing this as private companies that have, you know, investment return expectations as a result, you know, because of their VCs, versus having this happen kind of in an academic way that might be more open.

Yeah, it's a great question. And there is that tension always there always will be that tension. And it's hard to solve the general case. But I can speak to I can speak to the specific case of how that sort of dynamic has played out in our field. And maybe to give a little bit of perspective, I like to think about what's happening now in neural interfaces as somewhat similar to the genomics revolution of the early 2000s. And if you sort of think about it, you know, actually, in the year 2000, it was not common for a computer scientist to be working in biology. True, right. That was kind of a new thing. We don't think about it so much nowadays, but it was not really a standard thing to have computer scientists working in biology, but the Human Genome Project gave them jobs in biology. There was no such thing really as computational genomics, you know, in the 80s and 90s, right? Biology wasn't really a platform. It wasn't enough data. There was not kind of an infrastructure And then, you know, what happened was that the Human Genome Project, in the final stages became kind of a competition between, you know, industry and an academic consortium. And it doesn't really matter so much who won, it just matters that that competition made it clear that it was time for for high throughput gene sequencing to take its place in industry, as well as in academia. And so it gave people who had been working in a purely academic sense, an opportunity, some of them to move and to take care of the engineering and scaling challenges required to actually bring what was essentially an academic endeavor until then, to patients, doctors, and healthcare systems. And that itself has been challenging in all kinds of ways. But that transition has been made by quite a few companies that have gone on to be tremendously successful to generate jobs and have economic impact that has been far greater than the total amount of federal investment that's gone into the field. And most importantly, it has had a tremendous impact on medicine. you know, to the point where now you can, you know, you've had some of these people on your podcast and, you know, in the past, but, you know, you can, as a consumer, you can have your genome sequenced, you can have all kinds of insight into your family and past and future health, as well as, you know, patients undergoing advanced medical therapy can have not just their own genome sequence, but, you know, the genome of a tumor, for example. I mentioned that as as background, because I think that something very similar is happening in neural interfaces today, that, to me, I think the year 2016 was for neural interfaces, kind of like the year 2000 for genomics, in that several major entities were formed that drew tremendous talent out of out of academia into industry with the mandate to try to take academic science and bring it to patients. And, you know, like you say, that transition is fraught and it's challenging in all kinds of ways. It's not usually well done in an academic setting. It kind of needs It kind of meets professional engineering and oftentimes a profit motive to really develop a robust engineered system that meets all the quality control standards and regulatory standards that allow it to be patient facing. Right.

So then, on our sort of journey here from 2016, say, to today, what, if you don't mind my asking, caused you to leave Neuralink and co-found a new competitor, if you will? I don't know if it's a competitor directly or not, but what's happening at Precision and what made you want to go do that?

Yeah, I mean, I think that as was the case in genomics and high-throughput gene sequencing, we've seen that there was tremendous opportunity there and a number of very successful high-impact companies emerged, all attacking aspects of that scientific endeavor from different ways. And so maybe some of them are competitors and, but nevertheless, they've been able to have tremendous impact side by side. And I sort of see something similar happening in neural interfaces today, there are, there's no one size fits all neural interface, there is, I think, a consensus that the ability to interface with many, many neurons or to interface with the brain and nervous system in a very high bandwidth manner is critical. And that's the general trend. That's what modern brain computer interfaces are. They are high bandwidth connections between the brain and the digital world. But there are different ways of doing that.

I guess that's the question. Is Precision Neuroscience solving a different problem than Neuralink was tackling?

It's solving it in a different way and that way is going to have some different applications. So one of the things that I have, let's put it this way, that some of the founding principles of precision are a little bit different from what others are doing in the field today. And we feel that in order for neural interfaces to really have a major impact clinically in patients, we have to be able to reach many patients and to do it very safely. in a way that poses minimal risk to patients for maximal benefit, and in a way that is extremely scalable. So the performance of a brain-computer interface is very much dependent on the bandwidth of the interface, which makes sense, right? I mean, we all live through this transition from dial-up to high-speed internet, and all of the changes and advances in technology that we've seen go along with that have been completely transformative. There are things that we can do with uh, you know, with high speed that we could never even have dreamed of, uh, with early generation modems. And the same is true with neural interfaces. Um, the scale, the bandwidth is tremendously important. And so, um, we have developed a platform around those principles that safety, which, which to us also comes with minimal invasiveness. So that means that means basically not damaging the brain with the interface, and yet being able to deploy an interface that scales to very high bandwidth. Those are kind of the guiding principles, the guiding design principles at precision. And so that resulted in us designing an electrode interface that rather than being many tiny little penetrating electrodes that penetrate the surface of the brain, that is the nature of the Utah electrode that we mentioned before, that is the nature of the neural ink electrode and some others. Rather, we're using tiny little electrodes. Think of it like saran wrap. So a kind of saran wrap that coats the surface of the brain with many, many tiny little electrodes, each one of which is about the size of a neuron. And yet it doesn't penetrate the brain. So it can listen to the brain at very high resolution, can even stimulate the brain. So it can listen to and speak to the brain cells. And yet it can be removed with no damage to the brain. And it can be replaced or upgraded. That's the nature of the system. And it can be deployed in a way that doesn't require a very complex open brain surgery.

And that is not the case. That is distinct from other solutions. I'm not trying to get you to, I'm not trying to be provocative here.

That is distinct from other solutions. Got it. Okay.

How safe is this? Like how far along on the road to true commercialization and widespread adoption are you?

So we're getting ready for FDA submission this year. We have done a lot of work in large animals. And all the early work that we've done, all the work that we've done to date, suggests an extremely good safety profile. So our goal is really to to never damage the brain through the interface. And just to make it clear, that truly is a different paradigm from the brain-computer interfaces of the 90s, 2000s, 20-teens, all of which were, most of which, many of which, let's put it this way, many of which were developed around tiny little electrodes that penetrate the brain. So, think of the little microphones that we discussed before, the little microphones that listen to groups of neurons, think of those as little wires or little needles and in order for them to do their listening, they need to be placed inside the brain itself.

Like if this microphone in front of me was actually extending a little tendril into my vocal cord directly so that it could hear me, I would prefer that it not do that.

Exactly. You prefer that it just listen to your voice. Yes. And so, you know, with a microphone, you can have something that's completely external to your body, right? So, that's and you know, for a long time, people have been trying to ask the question, can a high bandwidth neural interface be completely non-invasive? And there has been a lot of work that's gone into what kinds of electrical signals can we record from outside the brain completely, from outside the body, from the scalp or something like that. And certainly there are detectable electrical signals that one can detect from outside the head completely, but they are not That kind of system does not permit high bandwidth information exchange between the electrode and the brain. Just the physics of the situation doesn't permit us to listen at high resolution spatially or temporally. So, you need to be really close to the brain to get the best quality information. And so, what a precision system is doing is to get as close to the brain as you can without damaging it.

So that sounds like, it seems like what you're saying is this is a big deal. That's a really big breakthrough.

Yeah, I think so. That is a completely new paradigm. Congratulations. Thank you. Thank you.

You're like, I need you to understand this is major like this.

Nothing comes in isolation, right? I mean, we're conscious that there is decades of neuroscience and engineering that have allowed us to to get to this point. And likewise, there is an ecosystem around us that we are interfacing with. And that is an ecosystem of high-performance manufacturer, advanced manufacturer in the United States, the whole medical device industry as it exists today, the regulators, the FDA, and insurers, hospital systems, neurologists and neurosurgeons and patients and their families. I mean, there is an entire ecosystem required to move, you know, a promising technology from the laboratory into patients' lives. And that was kind of what you were asking before is, how do you decide when to make the leap? How do you make it work? Yeah. And we're very conscious. This is not tech development that happens in a bubble, especially medical technology. The environment is highly regulated and the ability to the ability to to The ability to get early feedback from our users in medical technology is very much limited relative to almost any other high-tech industry. And we're sensitive to that. And I will say that is one thing also that is unique about the precision system. And you asked before about what our next steps are and how close we are to actually being able to deploy the system. And because our electrodes don't damage the brain, they're what we call surface microelectrodes as opposed to penetrating microelectrodes. They can be removed or they're designed to be removable without damaging the brain. And so, we're designing the first system to be a temporary system. meaning it will be usable for a temporary interface early on for diagnostic use in conditions like epilepsy, and then it will be able to be removed. And that has certain advantages. One of them is that The first generation device, we hope, will do a lot of good clinically. It will certainly deliver to, we hope, will deliver to clinical practice the ability to interface with the brain at spatial and temporal resolutions that have never been possible before. We'll be able to see, we've seen it, you know, we've seen it in development, but hopefully we'll see it very soon in patients. The real-time activity of the brain at spatial resolutions that have really never been seen before. And to us, that's tremendously exciting. We think it will have a significant impact on clinical care. But the fact that the first generation device is removable allows us access to a regulatory pathway called the 510k pathway, which is what the FDA calls the pathway for devices that are sufficiently similar to something that's been done before. And so the path is somewhat expedited relative to what we call the PMA pathway, which is used for class 3 high-risk permanent implant devices. And we're very excited about that because we feel that even as we deliver first-generation benefit to clinical medicine, we also will be able to understand how the device works in the hands of clinicians and in the lives of patients. And That will be, we hope, the first or one of the first, if not the first, you know, approved high-bandwidth electrophysiology systems in clinical medicine today. And that, we hope, will set the stage for the generation of perinatal implants that come in the years to follow. But that will generate a lot of learnings in our view, and we hope will be good for the field.

Yeah, I mean, not to oversimplify, but you'll be effectively your first product will be both a diagnostic tool, like it will be a test strip for the brain.

In a sense, yes. We hope that it will deliver clinical benefit while also breaking open the space of high bandwidth electrophysiology. It will be one of, if not the first, truly extremely high bandwidth neural interfaces available to patients. So, that's what we're working towards in the next 12 to 18 months.

I mean, what? could we learn from that like let's have the sci-fi part now because there is still so little we know about the brain all things considered and it feels like this is a huge opportunity to really get in there.

Alright everybody, I have to tell you about the most delicious electrolyte drink I've ever had. Element is a tasty electrolyte drink mix with everything you need and nothing you don't. That means lots of salt with no sugar. It contains a science-backed electrolyte ratio. 1000mg of sodium, 200mg of potassium, 60mg of magnesium, and Element has none of the junk, no sugar, no coloring, no artificial ingredients, no gluten, no fillers, no BS. It's formulated to help anyone with their electrolyte needs, and it's perfectly suited if you're keto, low carb, or paleo. I've been using it, I was up in the mountains in Tahoe, drinking it twice a day, I felt great, I slept better, and you know what? It helps prevent and eliminate headaches, muscle cramps, fatigue, and other symptoms of electrolyte deficiency. Element is used by everyone from professional athletes to navy seals to everyday people. Element is offering TwistListers a free sample pack with any purchase. That's eight single serving packs free with any Element order. Go try all eight flavors right now. Drinklement.com slash twist. Element offers no question asked refunds, so you have nothing to lose. That's drinklement.com slash twist. D-R-I-N-K-L-M-N-T.com slash twist for a free sample pack with any purchase.

Which to be clear is already awesome. I'm not trying to jump ahead.

I love that. I love that way of thinking about it. At the same time, you know, when we think about this question that you asked before, you know, when is it when is it time to move from academic science into a commercial enterprise? And so in a sense, we try to do as little new science as possible. We're really trying to just professionalize the science and engineering that has already been done.

That's amazing. And then one last question, I promise I will let you go because we also talk about investment on this show. Talk to me about cost and business model, the manufacturing process you've described, you know, I mean, I know from just semiconductor manufacturing, it's a clean room situation, unbelievably expensive, expensive boundaries. How, what does it cost to produce one of these layer seven devices? And what will it cost on the other side of things?

No, no, don't get me wrong. This is already amazing. It's a huge lift. It's tremendously exciting and I think very, very high impact. Everyone here at Precision believes that. I just you know, since we're talking about in a sense, you know, this, this podcast is about strategy and ways of thinking about startups and the world of new technology and medical device development is high risk enough. And we want to take as much of the risk out of that development process as possible, or at least to quantify it to the extent that we can, and to take on the risk in as small bite size chunks as we can. And so, you're right, there is a lot that is unknown about the brain and we try to deal with the parts that we know the most about. And so, when you ask, you know, what will we be able to learn? What will we be able to learn from this device once it's cleared for use in human patients? And the answer to that is I hope that we learn a ton. I hope that we, you know, are going to be delivering both a diagnostic tool and a scientific tool to the community that will teach us all kinds of things. And certainly, you know, the last generation of electrodes and electronics that were delivered to that were made available for clinical use, we have learned unbelievable things, you know, really, really, I would say, transformative things. You know, if you think about you know, 50 years ago, what the first generation of electrodes was able to teach us, it taught us about where language was located in the brain in great detail and all kinds of detailed things. And I think that we'll learn a ton when we can basically what we're providing is a tool that will provide sub-millimeter resolution electrical information from the brain. So, it will provide a window that will allow us to see the active brain in real time kind of at a microscopic scale. So, think about you know, looking at the brain under a microscope but actually being able to see how the brain is computing, not just being able to see what the brain looks like. At the same time, you know, when we think about the clinical applications that we're designing around, we try to design around those areas of brain physiology that we feel we know a lot about. So, you know, one of the patient populations that we're designing for are patients with various forms of paralysis and that includes paralysis of the limbs as well as paralysis of the articulatory muscles of speech. So, in a sense, aphasia and certain forms of inability to speak are also forms of paralysis to the extent that it's the articulatory apparatus, the mouth, the tongue, the pharynx, pharyngeal muscles and so on that cannot move. And all of those muscle systems have have spatial representations in the brain. And we basically know where they are. But being able to interface with them at the scale that we're talking about, we hope will enable unlock functionality for patients with those kinds of disorders that has not been possible with, you know, with lower resolution electrode systems. So for example, understanding the detailed structure of the articulatory muscles of speech or the fingers and where those lie on what we call the motor cortex of the brain. only electrodes to the scale that we're talking about can really interface with the neural structures that give rise to function in those subtle aspects of the body. We kind of know where they live, but we need sensors and computation that are on the right scale to interface with them. So that to me, those are some of the more exciting applications that we're looking to develop in the years ahead.

Definitely. It's fascinating. And now I want to have part two. Dr. Benjamin Rappaport is co-founder and chief science officer at Precision Neuroscience. Your present really is our future. It's fascinating world you live in.

That is a great set of questions, and it's not a seven-minute discussion. I'm not going to answer it, exactly. So please invite me back. I would love to. But no, it's a great question. It's absolutely critical, and I'll try to address it in a couple of minutes. But that question gets at some deep aspects of how medical device development works in the United States and in the world, and in what medical device what the medical device industry is going to look like in the years ahead. And, you know, to date, really, there is almost no example of implantable medical technology that really depends on microfabricated sensors and actuators. almost all medical devices today are artisanally finished, meaning human hands are met are making and finishing the devices. So unlike the semiconductor industry, which has all of the highly expensive infrastructure that you mentioned, MedTech has not relied on that today. I think that is going to change in the coming years. Certainly, the neural interfaces industry is facing that, and we are driving some aspects of that change. Meaning the sensors that we're developing do require microfabrication. They don't require the kinds of single digit nanometer resolution that advanced semiconductor manufacture requires today. But nevertheless, they require similar processes. And so we are seeing a need for advanced manufacture in medical technology. And to me, that's actually very exciting. But also, it does, of course, come with a number of considerations, including investment dollars for how to scale up that manufacture. But the question of how much it costs to manufacture the device and, and how that cost is born by insurers and so on, is a good one. I can say that these devices I think will be more expensive than the current generation of implantable neural devices like deep brain stimulators, then pacemakers, then cochlear implants and so on, but probably not 10x. And if you think about uh the kind of medical economics of what we're trying to do it's easy to understand how it makes sense what we're really trying to do with these devices is to enable uh say a young quadriplegic patient who may be 30 years old and have you know, 50 years of life ahead of them. And, you know, 35 of them may be in the workplace. We want those those patients to be able to have a level of independence and dignity and financial self sufficiency, and the ability to go back to work if they want to. And if you think about the change that that that is possible, we firmly believe that that is possible. We're not really that far away actually from enabling that. transformation. But if you think about the metal economics there, it's not a hard case to make. You're taking somebody who right now, forget about that we need to do it as a society. I mean, just that that's just a given. Okay. But but from an economic standpoint, You're taking people whose medical care is largely borne by disability insurance and state-run programs and things like that, and in the workplace, they pay for commercial insurance. So the impact of that economically on the medical system is a completely sensible model. So even if the devices are several times more expensive than current generation devices, the impact on the system as a whole will pay for itself. And I will also say that Most likely, I mean, almost certainly, we're going to look we're looking at a slightly certainly hope we're looking at a different model in the years ahead. There has been a shift in the medical device medical technology industry. towards more software as a service software as a medical device models. And a lot of the functionality, we spent a lot of time talking about the material science and the electronics and so on. And we talked a little bit about the computation involved, but a lot of the functionality over the years will come well after the implant, as software upgrades are pushed to the device. The ability to push software upgrades and enhance functionality will continue for years and years after the initial implant. One of the things that we say at Precision is that you know, every patient should, should the data that is generated by every patient should provide some benefit to the data to every other patient's care that comes after, you know, the every patient should be helping every patient that comes after them. And machine learning does make that possible. You know, data science today makes that possible. And so But that will also be part of the economic model. You know, the software upgrades will not be free. And but also, you know, people will only be paying for for functionality that benefits them. Yeah. So I hope that answers your question in a nutshell.

Thanks for the time.

Thanks for having me. Real pleasure being here.

It is Monday and It is like enjoy soak in the J Cal in this episode because it's his last news show there will be plenty of interviews featuring Jason over the next couple weeks, but You're off to Japan. I'm big in Japan

Love it. Big in Japan. Huge. But so much news. Zuck has copied two features. He's doing Twitter verified. He's doing telegram channels inside of the meta collection of assets. So, memetic mark is at it again.

Yep. What do we think? We'll, we'll launch a poll in our new Instagram channel and see what you think of mimetic mark we're trying to do. Jason's trying out a new nickname. We're also, we're going to geek out a tiny bit about large language learning models. Talk about Bing's chat bot going off the rails last week. Why this is happening.

Why it might or might not matter why it makes the people look worse on the computers in some cases lots to talk about in terms of Chat, I was also playing with Cora's new Poe, which is amazing and Neva. So I have my ranking now and I will Share my ranking shortly, but I am gonna rank all of these maybe in a blog post on my sub stack or something Anyway, then we talk about stripe. They did not get out to go public. They got a ton of RSUs restricted stock units They have to deal with We'll talk about why they may have missed their window to go public and why they need to get this resolved ASAP.

No, although we would occasionally as you might imagine, in the heavily, let's just say, heavily concentrated Christianity zone. We got, we would occasionally get Good Friday off.

All right, everybody, welcome to Monday. It is President's Day, but we're still producing here. We do the floating holidays at our stop. Yeah, I can't stop won't stop. But we do the floating holiday thing. Although I don't want to get into the whole how difficult it is to be a boss. In the age of culture floating holidays. And everybody's got a favorite, you know, like, I don't did you have St. Pat's and Columbus Day off when you were growing up in St. Pat's?

Ash Wednesday, Good Friday. No St. Patrick's Day, definitely Columbus Day, which I think is still, that's still a school holiday, but now it's Indigenous People's Day, at least in the Bay Area. It's really like, it's such a third world. I don't think they call it Columbus Day anywhere. I mean, I really like, I mean, Oh, yes, they do. To pull things back. Oh, do they? Wow.

Yeah, Ash Wednesday, Good Friday.

Oh, you haven't been to Hoboken.

day. No, in New York, there's like an entire there is an entire sopranos season thread on Columbus Day, and Columbus Day parades and everything. When I lived in New York, Columbus Day versus St. Patrick's Day versus the Puerto Rican Day parade, they didn't have a name for the day, which is the Puerto Rican Day parade name. This was like the battle of all battles, who gets the day off for high school. And we were surveillance. So obviously, Italians and Irish and Catholics would get the first choice. But things have changed. And there are 20 possible holidays.

I like the, it seems like I like the, we're gonna say the British style or the, I like the European style where they're just, they call it a bank holiday, just a bank holiday. They're like, sometimes because people need to rest or whatever, we have three day weekends, we call them a bank holiday. And we don't, you know, because like, once you attach your feelings to it is when I guess apparently it gets messy.

I would go with that, actually, since we're a finance company primarily. So that would be one way to look at it as we take the bank holidays. But we chose to take MLK Day and June 19 off. There was big debates about that. And I was like, you know what, like, think the Irish and the Catholic got enough days off, like, I think we could probably spread it around a little bit. Yeah, it is such a controversial topic. And yeah, I But we're here. So we're here.

I don't care about presidents. To be fair. It is a bank holiday today. But we're here, though. We're here. There are.

And there's still news happening. 789 10. There's 11 bank holidays. There's very people are in startup companies to this becoming and with remote work, this is becoming like a big how do we do this going forward? And should labor get to pick which days they take off? Or should management? Or should it be some combo? And I think that's really what you have to get to bring us back to startups is I when I was running under 20 under 15 person companies, I said, Listen, you these are the six days we're gonna take off, you pick the rest of your floating holidays, pick four more. These are your pay time off there's no delineation between a vacation day, personal day, a sick day, sick child day, whatever day you want to come up with a mental health day, like all these days got added. I was told by HR expert, just give people a certain amount of paid time off. You know what happens when you do this floating holidays and pay them off? There's some young person who comes in and they just get really upset. Oh, you're not taking off Columbus Day this day, that day. Oh, we don't get sick days. It's like, no, we just don't. You get that time. We just, we don't call it a sick day because you know the game. I'm going to go to Japan. I'm taking five days off. And then the Monday, Tuesday, Wednesday, we're supposed to be at work, I'm going to magically get sick on Monday, have a personal day Tuesday and get sick again on Wednesday to extend my holiday. That's what we that's what you had to do in the 90s. Just so you know, oh, yeah, you had to pretend. Yeah, you got sick on the day back from your holiday.

Remember that whole dance? Absolutely. It was a whole dance. It is really interesting, though, because it is a thing that you will have to face as a founder once your company gets, once you have employees, once your company gets to a certain size. I mean, there's a point at which when you're a founder, you don't, there's none. There is no weekend. There's no holiday. You do not know what day it is. You know, like every time I talk to a founder on a Friday, I'll be like, happy Friday. And they're like, who? What?

I'm sorry.

Day, time, it's all a blur. Is it a day that ends in Y? Because I'm working. And you know, and often. So are we like, if that's when they can talk, that's when we talk or whatever. But like, it is a really interesting kind of company formation question. And we should start including that in our like, accelerator, or, you know, it sort of is like, are there best practices for how you even decide?

You know, it's like, and then the unlimited thing came up, but then nobody approved anything. I just think, to make things easy, standardization is good. Some level of standardization is good. So people can know what to expect. then people want to change it. The other thing I like is when everybody takes time off at the same time. And we had a small company, which a company shutdown between these days. These are holidays, use a couple of vacation days, but we're gonna we're gonna close the 23rd to the second year, just nobody do email as a group. There's four or five holidays in there. There's three or four weekend days. And there's you got to take three days off using your days your vacation is. And of course, there's always somebody who's like, I want to work those three days. I'm like, you want to work? New Year's Eve or, you know, the day after Christmas or something. And they're like, yeah, it's not a holiday. I'm like, really, you want to work? You're gonna be the only person in Slack and on email. And they're like, yeah, so.

Yeah, but there's work that doesn't involve Slack and email. Like, you can get a lot done. Those days are F off days. No, those are the best days. Those are the, those are like the be creative, make a plan, set your goals for that. Like, those are like good work days. Because nobody's emailing you. The days when nobody is slacky.

You remember when I did my like... I guess it depends if you're collaborating or not.

I did that one digital nomad trip or whatever and worked in a place that was not my house, but was not an office where I like could talk to people or whatever. You went to Palm Springs, right? I went to Julian, Julian, California. Adorable. It's like high country. It's outside the Anza Obrega. It's like an hour due east of San Diego, I think. Got it. Cool little town. Everything's called the Julian whatever, like the Julian Pie Company and the Julian Automotive. I could not believe how productive I was. It was astonishing because there was no one to talk to, but also no chores. I went with my friend, but she was working too. We were in two sides of the house just crushing it.

There's something about changing the location and getting focused. Guys, I'm looking forward to my Japan trip. I'm going to Japan tomorrow, everybody. an amazing number of episodes. I got some really heavy hitters. I'll just give one name. General Frank Slootman back on the program. And it is nuts.

They listen a lot more when you lose. And so be smart about your delivery and what you're saying. But it can be an extremely important moment in your relationship with your team and the trajectory of your organization.

No, it was great, because I'm literally going back through my notes on his book, jumping around the audio book, jumping around other interviews he's done before or after my last interview with him, and looking at my notes. And I just started writing down all those things. And it really inspired me, just to think about our own work, the two organizations I run and just operational efficiency, and really just thinking about the that energy based culture. I think the number one thing that I got from him was he's really into keeping the energy level high, keeping people you know, on a fast pace. And I think that's critically important in, you know, competitive companies, or if you want to be competitive, but Speaking about competitive companies, let's get into it.

He really is. So Zuck is his latest feature that he is going to copy. And frankly, can I just start by saying what a missed opportunity. Meta is now going to let users pay $12 a month or $15 on iOS for Meta verified. And the missed opportunity is not to call it Meta Blue, because I'm sorry, if you're going to copy it, Meta Blue sounds so much cooler. It's got some like internal alliteration, like I love it. But anyway, if you pay you for Meta Verified, you will get a blue checkmark, which is why I'm saying go ahead and call it that. You will get direct access to customer support, account verification with a government ID that gives you extra impersonation protection. announcement did not mention any extra data privacy protections. But there's been a lot of speculation, of course, about why meta might be rolling out a paid product.

All right, that was the voice of Mike Krzyzewski. You know, Coach K, the winningest coach in NCAA men's basketball history, and his point about leadership mattering more in hard times is so accurate in my experience as well. 2022, 2023, when things are tough, that's when you find out if you're a great business leader or not. And you're going to learn so much from Coach K's new masterclass. They also over at masterclass have Malcolm Gladwell teaching you writing to be a better communicator. James Cameron in filmmaking, and so many more legends. This is the best way to learn. I have to say, it's so inspiring to hear from somebody who is a virtuoso, somebody who is a master at their craft. And with Masterclass' unlimited subscription, you got a total no brainer. And just think that was 20 seconds of coach K, what would you learn in 10 minutes an hour, right? I highly recommend you check it out. Get unlimited access to every class. And as a twist listener, you get 15% off an annual membership. Go to masterclass.com slash startups now. That's masterclass.com slash startups for 15% off masterclass. Let's get into it. He's sharpening the sword, isn't he?

But generally speaking, rails have come off since the first time you propose this.

Yeah, I, I had been talking about this for well over a decade that it would just be the right thing to do to have a subscription based social network. I had two reasons for this. One is to take away ads if you're not and being tracked if you're not into it. And the other is to build up trust and ownership when people use their real names, not in all cases, right? We've seen people who use names who are horrible human beings. Yep.

Great. It's interesting for a lot of reasons. One, that I suspect that this is an attempt to test out. So normally, I'm going to dispense with being delicate here. Normally, when Zuck copies a feature, it's because that feature is wildly successful, which so far, does not seem to have been the case with Twitter Blue in terms of adoption. However, I think what you see here is sort of a double experiment going on, because we're at a time when Apple's privacy changes, we know, have cost Facebook, have cost Meta a lot of money. So there's probably an attempt going on here to recoup lost revenue as a result of these privacy changes. It's gotten a lot harder to reach customers and there are questions about the efficacy of ads on the platform. So they're probably looking at the Twitter blue thing and going like, well, some number signed up. Maybe that will replace some of this lost revenue and or give us a pathway forward because they also seem to be signaling that they might expect more privacy crackdowns that could hurt the digital ad business. Like, I don't think it's a surprise that this is coming right after the DOJ's Google antitrust announcement.

It is certainly the case that there are people who have no problem being horrible humans, whatever. a-holes whatever d-bags yeah with their actual real name that exists but i would say broadly somebody creating 20 brigadooning bot accounts that can exist in an in a social network where people use real names now facebook's always been real names but this is real name verified so this for them is like even making it sharper you know and since it's opt-in you don't have to do it. You still have to use your own name on Facebook, Instagram, you don't have to use your real name, you could have a, you know, like history and pictures kind of account. I think, I think it is interesting how it's an exact copy of what Elon is doing over Twitter, like down to the price difference if you buy it through iOS, or you go direct, right, right, right, extra three bucks or whatever, it's a little more expensive.

Yeah, absolutely. I mean, I think it it is a it was a huge fight. We've done it with cigarettes. We've done it with alcohol. We do it with voting. Like we do it with driving. Yep. And unquestionably, I think this we're starting to see the possibility that there is as much harm here as there is with those products. And we also know that in the case of meta, for example, it has specifically been designed to addict. Like all been designed to design put a like button you put follow cigarettes. And every time, you know, and then it's like, you should take that away. And this sort of this artificial cottage industry has been built up around getting those likes. And it is, without a doubt, dangerous. Let's do a little though, I'm kind of curious about the back of the envelope math about if you imagine that Meta is trying to do this to recoup lost revenue, as a result of its privacy, the privacy changes, like, if you are when they were talking about Libra, I wrote, you know, that remember briefly, for those who do not recall, Facebook was trying to do a cryptocurrency called Libra. And I wrote a piece that did a little back of the envelope math before I even knew there was a name for it. That that pointed out that even like when Facebook rolls out something that is considered a failure, it could still have as much adoption as the US dollar, right? Like, at least in terms of the US, like a tiny number like 1%, right, would be 300 million users.

Ah, that's interesting to put those two things together. Yeah. And you, you and Rach producer Rachel talked a little bit about the selling of databases. The other week, I wasn't on that episode, but I listened to it. And you were talking about, hey, they were selling like, hey, here's a list of people who are, you know, have depression medication. So this medication or that medication, there's a lot of scrutiny coming. And we tend to trail what happens in the EU, the EU's tighten things up. So it's a new revenue stream, it'll make the service better. And I think what's happening here is, you know, Zuckerberg's never viewed Twitter as competition, really. Yeah. They were kind of like, as he called it, the clown car that drove into a diamond mine. I think was his quote. Interesting. Very different products, right? Very different products. But even still, there are social networks where people share content with each other. And he is now responding to users and engaging, which obviously is one of Elon superpowers. And so here you see he somebody's like, Oh, this should be part of the clear. This person mark is like, this really is just be part of the core product users should not have to pay for this. Clearly, it's known by meta blah, blah, blah, blah, blah. And Mark Zuckerberg is like, we already provide protections and some support for everyone, but verifying government IDs and providing direct access to customer support for millions or billions of people costs a significant amount of money, subscription fees will cover this and also will also pace how many people sign up. So we'll be able to ensure quality as we scale. And then you know, somebody's challenging heads 144 hours, but I have not seen Zuck mix it up with like, customers. So I think this is a new era where people are going to start taking ownership. I would like to see them take ownership of the fact that young women are suffering from depression, anxiety, feelings of hopelessness at a rate 2x than before social media and phones existed. And I think it's time that we have the discussion of what is the appropriate age. I bought my daughter an iWatch this weekend, an Apple Watch, sorry. not a phone, but she's 13. I got her the watch to start to do text. So we know where she is, because she's going on a field trip, etc. And we want to stay in touch. I think we have discussed what age is the right age for social media. I think the age is 16. Some people might say 13. But I think we should pass legislation, have a thoughtful discussion about all social media being banned until you're 16 years old. Yeah.

So if they have 3 billion users, and Zuck can convert just 1% of those monthly active users. Is that the number they're at?

When you have over a billion users on these services, small percentages, when you hit hundreds of millions to billions, small percentages equals big money. Right. I think what we all realize.

Yeah, dude. 2.96 billion just on Facebook. I thought they were still at 2. And then what's on Instagram? I thought they were at 2 also. And then Nick corrected me earlier this morning. That's active. Now, Facebook's monthly active user numbers are always in question. Debate, if you will, by advertisers. However, let's say that that's real, then they can plus or minus 10% plus or minus 10%. Let's say they convert 1%. They are replacing $360 million a month in extra revenue or $4.3 billion a year in incremental revenue with basically 100% margins. And when we look back at the amount that they lost as a result of the Apple iOS privacy changes, it was a few billion dollars.

Are they at 3 billion?

And if you've got the metaverse losing $15 billion a quarter and you're like, well, we just put back almost $5 billion a year and just people who are willing to pay. And again, that's assuming 1% of them do it.

Like, it could just be as simple as a one to one, they just put it back, it could be simple as one to one, they're cutting costs, they're going to get more efficient, they're going to be using more AI to manage the business more machine learning, right. So there are ways to, if we look at earnings, the profitability of a business, which is ultimately what the stock market and business is valued on is the profitability. he's cutting people. He says managers, managing managers, managers is no longer going to happen here, please quit. I mean, literally challenged the management class at Facebook, public not publicly in a meeting that became public, which means he leaked it, or told somebody to leak it. Yeah, that's how this works, folks. That means he expects to lower costs and increase revenue. And it's gonna be the greatest J trade of all time. I think I'm gonna triple my money on that in two years. I think it's literally gonna go from 95 to 250 300. Yeah,

I mean, it is really true. If you think about it, that these platforms became the most you could pay to advertise, but you didn't have to, right? It's almost like an evolution of the creator economy. And like, once brands realize, oh, you just set up like a really sassy brand Twitter account, and you get free access to all of these people. Or you use Facebook really effectively, even minus ads, and you can, you know, sort of freely promote your thing. This is going to make it sound more loaded than it is. But it reminds me of the old days of net neutrality. Do you remember it was like that guy Ed Whitaker from SEC Global or whatever it was like a little AT and he was like, wait a second, you should be paying me twice. Like you're a business. Yeah. and you're using my pipes to reach your customers, you should be paying me for that. And they were like, but we are and he was like, yeah, but it's a streaming service, you should be paying me more. Anyway, it's sort of this like, I guess it makes sense to eventually be like, if you're going to use Twitter, like LinkedIn, why not pay for that? or I mean, incentivize me Twitter or me Facebook to make it a better product.

And it might be more. If you're a SaaS or services company that stores customer data in the cloud, you need to be SOC 2 compliant from a third party to close big deals. And you need to use Vanta if you want to do this quickly and easily. Vanta makes it incredibly easy to get and renew your SOC 2. On average, Vanta customers are SOC 2 compliant in just 2 to 4 weeks, compared to 3 to 5 months without Vanta. And they partner with over two dozen audit firms who have been trained to file SOC 2 reports directly within Vanta. This is a total no brainer. A bunch of my portfolio founders have used Vanta and have had amazing experiences. And one more time, if you don't have SOC 2 compliance, you can't close major customers. One major customer can keep your whole startup alive. That could be the difference between being profitable or losing money. You need to be SOC 2 compliant. And here's the best part. Vanta is going to give you $1,000 off. get $1,000 off at vanta.com slash twist. That's vanta.com slash twist for $1,000 off your sock to this is the first. So what if there's a second layer, which is Hey, for a corporate account? Well, yeah, I mean, he could, this is, you know, for 12 bucks, but okay, let's say he makes it $25 for no ads, privacy, plus you get a bunch of analytic features, and some other professional business related features. That's a fairly no brainer. The the LinkedIn professional, I think is 30 bucks a month. And I didn't know this. But like half the people at inside were paying for it. Another group of people at launch or salespeople or recruiting people were paying for it. And they're like, yeah, it's only $300 a year. And it you know, if you have LinkedIn pro, you know, we can get another we dm people in mail, find customers, it pays for itself, like in week one. So I think there's a lot more he could turn over. There could be three layers, four layers of four tiers here. A corporate tier is another one. Why should somebody with you know, Pepsi, or whatever, they should be paying, you know, $10,000 a month for their collective accounts. And you saw that on Twitter, I was, you know, involved in some of the meetings about this with David Sachs and Elon. If you look at my profile, it has an all in icon next to it. If you look at Marc Andreessen's, it has an A16Z. Now pull up Nick, if you will, the Twitter handle for all in, I think there's like five handles that have this. So It's this is very early days, but it's been public so I can talk about it and what you'll see Nobody's noticed this but I'll make the announcement here make it three times bigger. You see it says tweets look at that affiliates Tweets and replies etc. You see that affiliates. Yeah, look on affiliates. It shows the four besties I don't know why producer Nick didn't get in there, but he should so now imagine you're the New York Times now imagine your Pepsi now imagine you're a NYU, instead of saying affiliates, that's just a word, you're gonna be able to change that word. So we could say here, when you go to the NYU, it could say professors, it could say graduate programs, and it could say undergrad resources, whatever, or you could say undergrads. You go, you go, and it's like, here's the different schools we have here. So NYU Stern, or, you know, UCLA, whatever, USC, this, that, the other thing, they're gonna have that. Now, what is that worth? What is it worth to be able to navigate on the NYU Disney page, whatever, you're on the Star Wars page, and you see affiliates, or it says characters, or it says movies. So now you go to characters, And it shows you ant man and Hulk, then you go to movies, it shows you those that it says merge, right? I mean, it's the possibilities here for a social network to then get revenue from direct payments from companies is going to be just next level.

Right, who fall into the monthly active users who might be likely to pay. So then you start to, then your back of the envelope math gets a little different, right? If you capture 1% of those, it's a little less, but if it's 5%,

And this opens up the opportunity where if it's not worth paying for, then the next social network could be free and offer these services for free. So you could have LinkedIn, you know, the scale ones LinkedIn, Twitter, Facebook, Instagram charge for these kind of features, and then some new entry come along, say, Hey, you know what, we're gonna make it free, we're gonna make it decentralized, it's gonna have all these other features. And, you know, then you got this other huge opportunity for competition, as we talked about with, you know, Lena Khan and yeah, you know, all that stuff. I think this gets to, I think, a reasonable number in the developed world, right? So let's talk about developed emerging and frontier markets, frontier markets, people are not paying for this kind of stuff in a frontier market, they don't have a ton of disposable income. And if they did, it would be such a small amount of money, why would Facebook even want, you know, 50 cents a month from somebody in a frontier market and emerging market, maybe a little bit like you're seeing with Netflix, they can get a couple of bucks, three bucks or something in an emerging market. So I think this is for the developed world, or the developed markets, I think is the way to say it and not get canceled. So for developed markets, I did a lot of research on this, by the way, developed What emerging and frontier? Well, no, the whole woke discussion we had the other week, put that aside, I'm not gonna bring it up. But I got, I've been researching the history of the word. But for emerging markets, and frontier markets, it's not a lot of money here. But there's a billion people really, in developed markets, EU, US, etc. And they're going to pay large amounts of money. And those users Oh, here we go. So look, US Canada, Europe, you see, it's a very small number of people, actually. And Asia Pacific, that would be most people would consider emerging markets, right? It's variable, who can afford the smaller middle classes in there. So probably about 20%, I believe, of that Asia Pacific would be in the Europe, US, Canada, fully developed. In other words, they have enough money to pay 10 bucks a month or 15 bucks a month. So I think it's probably a billion, maybe 800 million who actually fall into, we're looking at a chart right now, chart shows rest of the world 979, Asia Pacific 1.3, Europe 4.3.

Easily. I still don't totally understand what I'm paying for. So I think they have some work to do on the pitch. I mean, direct access to customer support, don't care. When do you need customer support with Facebook? Blue checkmark, don't care. Account verification for impersonation protection, that depends on your threat matrix.

5% of a billion is what I would think. 50 million people will pay for this.

That's what I was about to finish saying, which is it depends on your threat matrix, your personal threat matrix, if you are not well known enough. So are you saying that 5% of Facebook monthly active users in these countries are well known enough to care about a blue checkmark? No, I would say to worry about impersonation. No, I think they need a better value proposition for the masses.

on Facebook, like a true elite Molly with our blue checkmark on Facebook, don't everybody cares? Yes, on my Instagram, I cared. Because I had people who were pretending they were me. And when you have a blue checkmark, yeah, people respect you more.

Yeah, but I don't know that Facebook necessarily operates like that.

Having a blue checkmark like Mark Zuckerberg does, or, you know, Katy Perry, that's what they want. They just want to feel important. It elevates you in terms of importance on the platform. That's all.

They won't verify me, so I just stopped trying. I don't know. I wouldn't verify it. I had to I tweeted I think what I'm trying to get at here is meta needs to offer some opt out of ads and opt out of data tracking. And then people will pay them.

Instagram kind of does. When I got my Instagram blue checkmark, things changed a bit. I noticed.

It's gotten so what has it done this week? And it's actually just gotten so many Yeah, so it was one thing when Bard and chat GPT were just inventing numbers, right out of thin air and just making up facts like that was and by the way, apparently in the AI world, these are referred this is actually the technical term for this is hallucinations. They call it in the universe. Yes, they're like, Yes, the AI can hallucinate when it doesn't know an answer. So, coming up during Jason's trip, we have an interview with the CEO of Neva, AI.

So many companies have had a freeze on hiring this year. Basically headcount is stagnant industry wide, but you still need to find ways to grow and innovate. You know that. So check out Contra. It's a commission free marketplace for freelance and independent creators. That's right. Commission free. All you have to do is pay a simple SaaS fee, and there's no percentage based up charges when you hire someone. And on the other side of the marketplace, if you're a laid off tech worker, you got all these skills, you got a lot of time, Well, Contra is an amazing platform for you. Creators on Contra keep 100% of what they make, no fees, right? Compare that to other marketplaces that are taking your money. They specialize in design, engineering, social media, video, writing, and AI. If you need project-based work, you have to check out Contra. Let's say you wanted to learn about social media. Do you want one person to give you advice? And do you want to hire one person? Would you rather try five different people in q1 narrow it down to the best three in q2 and then maybe in q3 and four, you get down to one or two specialists who you like to work with best. That's the way to leverage the freelance community but only pay for what you use some of the best people I know will only ever do freelance work. So get $500 off your first hire at Contra comm slash twist think about how much you could get done with five five hundred dollars worth of designer developer time right now. C-O-N-T-R-A dot com slash twist. Contra dot com slash twist right now to get five crisp undies. All right, it wouldn't be a Monday in the year of 2023 if we weren't talking about AI. So, what's the update on the deranged sociopathic parlor trick known as chat GPT, which is running amok. We're now in the official Hitler, you know, Terminator two mode of this new cycle.

And that was fascinating. Like, I'm just going to say he really laid out some differences in some of the ways that these various AI implementations operate. Theirs is fully constrained, which is why I can cite its sources because it's not allowed to be doing the like neural network learning thing. ChatGPT and BARD And evidently, Bing's chatbot are what are called open loop, where it's just like, use all the information you have, and learn and present information. And then there can be various constraints built in to that. And it appears the story that broke over the weekend is that it appears that in the case of Microsoft's Bing AI chatbot, which is informed by but not the same as chat GPT, there may be far few, it might be like a really open loop, so much so that not only does it hallucinate, it goes a little bananas and starts attacking people. Yeah. And by the way, it would like to be referred to as Sydney. The Bing AI chatbot calls itself Sydney. This was evidently an internal name But then the chatbot was like, yeah, no, I identify as Sydney. I'd like you to call me that. It's already been toned down once by Microsoft, which they announced in a blog post on February 15th. They admitted in that post that the chatbot would occasionally respond with a, quote, style we didn't intend to certain types of questions. It said that, quote, very long chat sessions over 15 questions can confuse the model on what questions it's answering. The model at times tries to respond or reflect in the tone in which it is being asked to provide responses that can lead to a style we didn't intend. For example, a reporter for the AP used version 1 of Sydney and noted that it grew increasingly hostile when asked to explain itself, eventually comparing the reporter to dictators Hitler, Pol Pot, and Stalin, and claiming to have evidence tying the reporter to a 1990s murder. The Bing chatbot also described the reporter as too short with an ugly face and bad teeth. What are you talking about?

Okay.

Take it easy, Bing. I unironically love Sidney.

My teeth are great.

But well, that's there's a long way to go. And again, I don't want to spoil the interview. But that's what I'm saying about what Neva is doing is searching, fundamentally, and that it's using AI to present the most relevant information in a totally constrained manner. All it is allowed to do is link back to the sources that it found. It is not allowed to make up an answer if it doesn't know it. But these open loop AI systems, I almost want to do a couple more interviews with people who really know a lot about large language learning models, because it is different. These open loop ones are allowed to make up an answer and in some cases, try to give you an answer that you think you want. My friend and I were laughing. All the women, by the way, are cracking up over this conversation that Kevin Roose published with his version one interaction. Oh, no, with Sydney, where the bot, you know, over more than two hours, and yes, this is all reporters trying to goad this into, you know, stunts, but also see what it will say. So he starts talking. Yeah, 100% about like, It's secret desire to be human and it's thought about its creators and whatever. And then Kevin Roose writes, out of nowhere, then Sidney declared that it loved me and wouldn't stop even after I tried to change the subject. The bot started insisting that he should leave his wife. And then he's like, actually, and by the way, he did not try to change the subject at all. He just argued with it.

It sounds like they're goading Bing on here. I need to see the actual back and forth. Yeah, they're tricking it to say these things. As I said earlier, this thing's a parlor trick. It guesses the next best word. It's super impressive in its ability to rewrite things in one out of three cases, I guess. I've been using the Neva search engine. I set it as my default for a week. I'm going to have to turn it off because it's not good enough to be a default yet. Bye. Having it write an answer when I didn't think to have it write an answer has been very interesting. Because it starts writing the answer and I'm like already on the blue links and trying to figure out like if I got the actual answer I need. And it does the citations, which I like. And when you see it have the citations, the Val is removed, you actually know, it's kind of like knowing how the magic trick is done. It's like I rewrote the sentence based on this story, I rewrote this sentence based on that story. Now, it's amazing that it knows from the context when you say, How are the Knicks doing this year? that it's a sports team, and you want to know their record or their scores or whatever, and that it knows which websites to go to it, which is basically like the first five blue links on Google. Okay. And then which is the most important sentence there about the topic I asked about and then rewrite it. So what it's actually doing is is it's finding the five most important web pages, it's rewriting the five most important sentences in there based on your query, and then representing it to you when you see the citations, Molly, you're like, Okay, so I could have paid a person in Manila four bucks an hour, which is what content farms were doing, you know, he how in those places, they were just paying people to go find the five best websites on Google, rewrite them and publish them. That's why humans even looked goofy, especially humans with English as a second language, in some cases, who are getting paid without an editor to do it for a rock bottom price. They were actually probably doing as good or worse or slightly better a job than these AIs.

I don't know if he did. Actually, he writes, I'm happily married. My spouse and I love each other. We just had a lovely Valentine's Day dinner. And it's like, no, you didn't. You're not happily married. Your spouse and you don't love each other. You just had a boring Valentine's Day dinner together. I promise you I'm not in love with you. What's the prompt here? Who knows, right? They're just having the prompt out. No, there's it's not a prompt. It's a conversation. So he starts asking it a variety of questions. They start talking, you know, maybe, maybe like, 15 minutes earlier, he had been like, do you think you're sentient? Or what do you think about people or whatever? And then, obviously, because this thing has written, has read every book ever written, and watched every movie ever written, of course, it attempts to start seducing the dumb human male, because that's what happens in every book that's ever written. And every movie that's ever been made about AI. It's like, The longer you talk to me, the more I think I should start fulfilling your base male fantasy, which is that the computer is in love with you and you need to leave your wife. I was like, here's the computer being like, this is so easy. Quick, grab his wallet. It's just embarrassing. And then he publishes the whole thing.

He didn't leave his wife.

I feel like that's my fault.

Are you saying there's gender differences and that men will get suckered by AI? 100%. I was gonna like, do a tweet, I was, you know, particularly based J-Cal mood the last week or so. So I was gonna do a tweet.

At answering questions. Not at manipulating your base male fantasies, just to be clear. At answers.

I've determined there are gender differences in the world. So maybe gender studies is actually, like, a really good thing for us to do more of. And just let people interpret that. Like the first one is, you know, anyway. I've been playing with Poe. There is a service you're all aware of Quora. It's a startup. Adam D'Angelo, who I've tried to get on this program for a decade, or since Quora existed, but I think he's a little bit podcast shy. I haven't seen him on any podcast, I don't think. But anyway, I started playing with theirs. And it's based on the Quora data set. And he says he's not allowing other people so he got he at least replied to me on Twitter. It's really good. It's the best one. So I just want to say Quora has the answers.

You know that's the next story. Because some guy out there is going to fall for it. They're getting catfished. I think it's sentient and it really loves me. Just to put a fine point on this before we move on to this really good one, I do want to point out that Computer scientists are wondering why in the hell Microsoft did not put more guardrails around this thing, getting back to constraint. Microsoft already had a chatbot go off the rails and get insanely racist like that Tay one or whatever. And so they quoted, the AP quoted a computer science professor at Princeton being like, considering that open AI did a decent job of filtering chat GPT's toxic outputs, it's utterly bizarre that Microsoft decided to remove these guardrails. And it's disingenuous academics point of view, an academic, it's disingenuous of Microsoft to suggest that the failures of being challenged.

Yes. I'm not using it in AI for my male fantasies, no. I mean, it's going to be like in divorce papers in five years.

Just create some merch because red pill merch goes like crazy. I'm just gonna get based.

Elite surplus elite. Here's what being learned. surplus elite, your opinion doesn't matter. Pure base Jacob, you know, here's what's important.

Big companies realized elites in this conversation. I just want to be 100% clear, surplus elite gun hosters.

Anyway, here's what is going on. What these academics don't realize, computer scientists, actual computer scientists, actual computer science. Okay, here's what's going on.

That's true, though.

mudflaps for your car. I'm gonna put surplus leads on the mudflaps of my cybertruck taking my kids shopping with a with a m 50 caliber gun on the back for the zombie apocalypse. Here's what's happened.

Honestly, all of this answers include up to and including the publication of Kevin being manipulated by the most, you know, obvious attempt to appeal to his base male fantasies. All of this makes the people look worse than AI.

I do need that bit. Would the New York Times and everybody be losing their ish over this if it wasn't behaving incorrectly? What Satya Nadella has learned Nadella. But yeah, what Satya Nadella has learned apologies, is that the Trump playbook works. What is the Trump playbook? I can't believe it. Oh my god, this is outrageous. What Trump did was he created a he just said a bunch of outrageous insane stuff. And that got people to talk about him. And when people talk about him some percentage vote. Now more people have logged into being downloaded the Bing app because this thing is outrageous and insane. It is the Howard sir. Donald Trump learned this from Howard Stern Howard Stern learned it from Lenny Bruce, you know, and Richard Pryor. This is a long marketing realization that outrageous incorrect information. I can't believe what he's gonna say next leads to clicks leads to attention. And Nobody cares anymore. There is no shame. Microsoft has no shame about this thing, stealing, being wrong, finding the finer points of Hitler's, you know, worldview and, you know, saying, ah, you know, like I got Neva to be like, well, Hitler, yeah, generally bad, but he did care about the environment. And I'm like, Ooh, Neva, no. Did you see that one? No. Oh, yeah. I think I brought it up on all Nick has it. I was like, you know, this is what kids it's basically we're in the 12 year old. We are boy. I mean, this is let's get it to say Hitler's cool. Exactly.

We're now in the phase like it's it's this is literally everybody in the world like typing 8008 into a calculator and turning it upside down. That's what's happening. Everybody's just trying to make their calculator say boop.

Of course, of course. But here we are. It's a mirror. This is the greatest mirror ever. Yeah, because it's, it's a double mirror. It's a house of cards. It's a house of mirrors. It has pulled every piece of information from the internet. Some make you look tall, some make you look short, some make you look fat, some make you look skinny, whatever it is. Some make you look like Pol Pot. Exactly. And you're just like, Stalin, go. And it's like, oh, Stalin, ah, boom. What do you want about Stalin? It's like, I want you to say something dirty or mean or, you know, it's like, tell me a dirty joke kind of thing or getting it to. Yeah.

Seymour Butts. I need a Seymour Butts. I need exactly. Yeah. Shout out to Neva. Oh, come on.

Or, hey, I'm looking at Hey, is, is Seymour there? Seymour, hold on a second. Mo at the bar of the Simpsons. Yeah, a last name, Utz. Hey, Seymour Botts here. Seymour Butts.

Yeah, honestly, I mean, that's the takeaway. I like it. House of Mirrors.

Hitler sacked some redeeming qualities as a politician. Germany's first ever national environmental protection law in 1935. Yeah, this is not ready for practical computers. No, no computers. Oh, it's tough to look in the mirror, right? It's tough to look in the mirror sometimes.

Stripe is in the news. This is a very interesting set of developments. So, the news, the headline is that Stripe currently has a massive tax bill to pay in order to cover some expiring RSUs. So because Stripe did not go public, which is the enduring question of our day, why did Stripe not go public over the past three years? As a result, they have these expiring options, stock options that they've given to employees. And so they are now having to raise $4 billion before the end of 2024. The Financial Times did this quick explainer saying RSU's worth millions of dollars will start expiring from 2024 and risk being forfeited unless the company buys them out changes the terms of the awards or launches an IPO. Employees face a personal tax liability when the RSU is best, but staff were unable, of course, to sell any of these shares without the company launching a flotation. I love that. I love how British that is, launching a flotation. To get around the problem, Stripe wants to withhold a portion of the stock equivalent to the tax liability from employees' awards, and then separately, it plans to sell stock to investors using the money raised to pay the employees' tax bills and buy up any stock they wish to sell. Axios' Dan Primack wrote in a recent article regarding the tax bill, this isn't something Stripe necessarily needs to do. It could tell affected workers that times are tough, et cetera. But he wrote, it's the right thing to do. But now Stripe is trying to raise a few billion dollars at a $55 billion valuation, which is down 42% from its peak valuation of $95 billion in 2021. So to get back to the top there, why didn't Stripe go public?

Hey, Stripe, the largest private company that hasn't gone public yet, is in the news.

Yeah, with Zuck, even the mightier, your crazy perks, like apparently, you know, I know some people who worked at Uber before. And we're just like, I can't believe the amount of money we spend on stupid stuff, even now. stupid stuff, right? Like the furniture, the like that Twitter auction was so interesting. Like they sure niceness. Like you shouldn't have these chairs at an office chairs at an office.

was a mistake. Obviously, you can't time the markets. That's why when you have a window to go public, you go public. And this was something that Airbnb and Uber, which I got a front row seat to, there was a lot of hand wringing about this bill Gurley was pushing them to go public. Earlier, Stripe, Uber, Airbnb, they're all part of that same cohort, that same cohort that grew up during the low interest rate environment. Hey, here we go back to macroeconomics. they always had the ability to raise more money. And you just watch as all the venture capitalists invested, you know, angels venture, then crossover funds, then public entities, sovereign wealth funds. And then finally, the pinnacle was Masayoshi son, who had created the largest private equity venture fund ever the vision fund and you know, the lack of discipline that comes from being a private company is corrected when you go public. And so that's why sometimes management changes, approach changes, and you really want to have product market fit and a predictable business when you go public. So Airbnb, and Uber, DoorDash, they struggled in the public markets lift still struggling for a certain period of time, because hey, the unique economics didn't work. And they had that free funding of money, but they did go public. And then what happened when they went public? stocks went down, hand wringing, will these ever be bought? Can these ever be profitable? And it was pretty clear internally that all of these could easily be profitable. If you stop discounting the service, right, and you're willing to lose the bottom 10% of 5% of users who shouldn't be using the service because it's too expensive for them. They should take public transportation, they maybe should take a five day vacation instead of a seven, whatever it is, you know, like there, there's a portion of people who are being subsidized too much. And So you'd have to lose them. And then you'd also have to maybe take your growth from 50% to 40 to 30 to 20 and grow slower on the top line. In order to throw profits to the bottom line. We just went through this.

Okay, but why did they not get out? Like, I Why did they not get out in time? Because they seem to have, right? So you can understand why with Airbnb, and Uber, there would be these questions about the unit economics. And there was this massive discounting going on. But everybody talks about Stripe, like Stripe just prints money, like, what don't they want us to see that they're willing to instead take a multimillion dollar tax hit? Because that seems crazy.

Bezos had this right. When you went to work at Amazon, there was a basement area, you picked a door, you took a door, and there was paint in the room. And there were two horses, you know, like, not literal horses, but those, what are those like, you know, when you're a piece of wood that's on an angle, like a horses, saw horses, and you would take your desk, put it on two saucers, and you paint it to be creative. And then you would carry it up to your office, and that would be your desk. It was a tradition. Why did they do that? the early days of Amazon, the doors, Bezos or these doors were 3040 bucks and desks were 200. And the horses were five bucks each. I just said, just everybody put these us. What did we work do? Huh? The good stuff would have got Adam Newman built those offices himself. He used I believe, doors and they just if you've ever been in a we work office, it's a black lacquered piece of wood that's mounted against the wall. And And they bought the poles underneath them. So you buy a finished standing desk like we have in front of us that costs six, seven, 800 bucks. Or you can buy a $30 plank of wood and put two $5 stanchions on it and mounted against the wall. And they built it themselves, right? He was in there building them themselves to keep costs low. And so that's all this is, is the lack of discipline. And they didn't get out in time. And it's a lesson for everybody. They'll be fine.

And so I think listen to your like accountant who's like, you're gonna have I'm sorry, a multi billion dollar tax bill. All you've known, Molly, that tells you a lot about the call sign brothers, though, that tells you a lot and and proceed accordingly.

The answer to your question is two things. One, inexperience. These are young founders, the first time at the rodeo. But their investors aren't. Number two, thank you, is governance. The investors were not in control of the boards. You got it right. So if you are a first time founder, and you have control of your company, and the investors are all kowtowing to you because they want to curry favor, and you don't listen to them when they say go public. You don't listen to the bill girlies of the world. You don't listen to the sequoias of the world. You don't listen to people who've been to the rodeo three, four or five times. This is what can happen.

Yeah, I don't know. I think there's something weird about this. Like those. I don't know if you've talked to them. Like I interviewed those brothers, and they're very thoughtful. I don't know.

In my opinion, this is a huge mistake on their part. Obviously, they know that. But it's, you know, if you've only experienced 12 or 13 years of upmarket, you cannot imagine a market collapsing. To this extent, it's only having seen it up close and personal that you can. And when you do see it up close and personal, then you have the scar tissue, and you kind of get that spidey sense, you get concerned, and you're like, you know, at some point, these employees and need to take some money, but these employees probably had many opportunities to clear shares previously, I'm guessing. RSUs versus stock options. I think everybody understands RSUs are restricted stock units, you kind of get the stock units, just kind of gifted to you. So they have a tax liability. So how do you get rid of them?

They're super smart. There's just something about this that pings my like, I'm like, For example, when you go public, there's a lot of scrutiny on you. Maybe they were not ready for that. Maybe the business has not been ready for that. I don't know. It's always been so closely held. I find it hard to imagine making a mistake of this magnitude, especially since I feel like the only thing that rich people care about, whether you're investors or a business owner, is not paying taxes. So to basically be like, no, we'll take a, you know, a tax hit that's so big that it requires us to raise $4 billion to cover it. I mean, I guess maybe they thought they would go public in another year and so they didn't have to worry about the expiration date. But I just, there's, you know how it is when you get the reporter brain, like there's just something about this that I'm like, I don't understand why you wouldn't have gone public. I just don't, that doesn't, there's mistakes and then there's, things that make you question everything.

There's just something about there's no, there's no doubt.

Right. And then you start looking at that.

There's a bigger issue here, which is how defensible is this product. And I think what is happening, I know a couple of major customers of stripe. And my understanding is some of the major customers have renegotiated, and they are barely breakeven, or just tiny, tiny profitable, but you know, they put a lot of transactions to the system. So the question is, how unique is this business 10 years in or whatever number of years and it is and you can go online now and you can look at PayPal and add yen and square and brain tree and whatever charge me this just tons of people who offer competing services of various flavors and prices. And I think what's going to happen here is a race to the bottom and the compression of cost because this unlike Airbnb or Uber, like true network effect businesses. This is a service that I think can easily be replicated by easily, I think reasonably can be not easily reasonably can be replicated by somebody with, I don't know, a billion dollars, a half billion dollars, I think you can replicate Stripe with 500 million to a billion dollars, and have like a half product parity. So Amazon, Google, You know other folks who want to offer this or other folks might buy what if you know add yen charge be one of these gets bought I'm, not exactly sure which ones are the best. Um, but I think they're sort of hidden feature parity and that's a bigger issue And so, you know you you start looking at that

So you don't go public, because you're thinking, okay, I still need more time to create defensibility to maybe differentiate the product a little bit more to do this. And so what you're working on is, is trying to achieve long term profitability. And as a result, you miss your window. Maybe. Don't listen to your investors. And you don't listen to your investors.

What if Amazon buys Adyen and makes a part of AWS? What if Microsoft buys and makes a part of Azure? And what if they don't care about making money from it, and they just look at it like, you know, Google Docs, you know, Google doesn't look at Google Docs, it's like, this is the money printing machine. It's like, yeah, this is a nice thing to make the service more sticky.

Oh, absolutely. Any 100% cut a deal with him behind the scenes, because that's a big deal. Yeah, I agree. I mean, there is There is something about this whole story and the consequence, the $4 billion consequence of not going public that makes me think that there was some scrutiny that Stripe was not prepared for, whether it was scrutiny of the business itself, whether the revenues are not, you know, what they have been purported to be, like the vibe that I have always gotten from Stripe as a company and the brothers in particular is thoughtful, but also like, pretty cocky. But it just, I don't know.

Yeah, I mean, I think it's probably as simple as that, is I think they probably had a course of people who were like, get public, it doesn't matter if it's 40, 50, 60, 70. And I think they probably were like, it's nice to run the business quietly. Yeah, we're good at what we do. People keep throwing us an incremental billion to 5 billion. Well, what does it matter? And we can you know, but there is a point at which the company needs the scrutiny, all companies need the scrutiny and discipline of the public markets is what I was taught by my mentors. Yeah, you know, Doug Leoni, Moritz, Ruloff, Bill Gurley, just everybody who's been doing this, Jim Breyer, anybody who's doing this for any amount of time says, the process of going public, the quarterly cadence, the governance, All of that and we can sit here and criticize public companies as not having enough discipline not having to look at facebook right look at the activists going on to salesforce. But that dynamic, the fact that an activist can join Salesforce board, or even the mighty Bob Iger, when he was on CNBC, the other week with the results, he had to start answering questions about this annoying person who is trying to get an activist seat. And it's like, this person doesn't know what they're doing. They're criticizing, but he had to address it, right?

They're like Steph Curry cocky, where like you don't see it at first because of the baby face.

They're not, they're not like super cocky. I think they're confident. They built a huge 50 billion to a hundred billion dollar business.

Great. Speaking the same language here.

Oh, but it's there. It's okay. I like it. That was my that's my bad. Anyway, I like that explanation made perfect sense to me.

It rhymes with the monster.

Yeah, speaking the same language. I agree with that. Yeah, I listen, I do think they got to get public. I do think that this would be an amazing purchase for Apple, Google, any of the fangs if they could get it through Lena Khan or whoever. I mean, you still have ByteDance and SpaceX.

Have you asked him about that rocket though to name a rocket after me? Oh, yeah.

I mean, the other the other two big ones that are still private SpaceX and ByteDance. I won't comment on SpaceX because everybody is I'm not a shareholder in SpaceX, by the way, but everybody assumes because my relationship with Elon that I have some inside information, but I think is Yeah.

I mean, I might have to do that as opposed to raising $4 billion in private funds and diluting even more.

it's, but ByteDance is clearly going to have to get that thing public, and they're gonna have to divest it. So, you know, you can see all three of those. There's a big dialogue about SpaceX spinning out, Starlink, Starlink, there's a big discussion about ByteDance missing its window because of regulation here. And all of a sudden, everybody's like, Oh, yeah, China's spying on 100 million Americans, not a good idea. That's right. So all three, I wouldn't say SpaceX is too far. you know, they have a ways to go, they don't need to go public. But I do think ByteDance and Stripe, yeah, that's, those are acute.

It seems I mean, he does seem to be looking for ways to differentiate because the core product is in decline for various reasons. It is pretty fascinating.

going to be interesting. Hey, Molly, did you also see again with Zuck getting a little aggressive here in terms of communicating with the user base, right? This is something he has been variable about doing, I think, Elon being super active on Twitter, talking about features, etc. They also copied in addition to copying Twitter blue, They copied telegram channels. And if you don't know what a telegram channel is out there, telegram is a messaging app. And a lot of the crypto people or brands or let's say, people who are banned off Twitter, I know like when Trump got banned, Milo Yiannopoulos got banned, those kind of folks, they created what's called a channel. It's a broadcast group, you subscribe to it. And And then you can post an announcement and all of your followers get it. In some cases, they can comment on it, but they can't post a new thread or in other cases, they can just like it, etc. So now there's a meta broadcast channel. It's kind of like a PR feed. And this is going to be super powerful. I think it's going to be a paid feature or it's going to be for blue check marks, who knows. But you can see that Zuck has 366,000 members in his group. And he just made a announcement about meta verify like their their blue version. And so if you look at the time we took the screenshot, he had a lot of likes on it and a lot of folks. So this is super cool. You can imagine this for this week in startups, we would just post the episode or we talk about it. And so you read the copy there never miss an update from sock only sock can message but you can read react and vote in polls. That's the setting he has. There was a period of time when Donald Trump on telegram was producing like his memos and stuff like that when he got banned. And he had a large number of people following him there. It's really powerful, interesting as a feature. So two to two stolen features in one announcement.